Categories
Salesforce

Salesforce Loops Acquisition: What It Means for Tableau Customers

Last updated: October 19, 2026

Short answer: On October 12, 2026, Salesforce signed a definitive agreement to acquire Loops, a small AI analytics company whose agent, Scout, watches business KPIs and explains why they move. Salesforce plans to fold the Loops technology and team into Tableau. The deal is expected to close in early November 2026, the price was not disclosed, and Salesforce has not yet said how the features will be packaged or priced. For most Tableau customers nothing changes today. The useful work now is getting your KPI definitions and data in shape so an explain-why agent has something reliable to explain.

The Salesforce Loops deal is small by Salesforce standards. Loops has fewer than 50 employees, and one unconfirmed estimate puts the price at around $30 million. Searches for it still jumped more than tenfold in the US over the past week, mostly from Tableau users and analytics leads asking the same two questions: what does Loops actually do, and will it change the way we use Tableau?

This guide covers what is confirmed about the Salesforce Loops acquisition, what is still unknown, and what Salesforce and Tableau customers can sensibly do before the deal closes.

What Salesforce announced about the Loops acquisition

Salesforce announced the agreement on October 12, 2026. The key facts from the Salesforce newsroom announcement and early coverage:

  • What Loops is: an AI-powered analytics platform that monitors business performance, explains why key metrics change and suggests where to improve, using machine learning models.
  • How it works: it has headless, agentic capabilities and can be triggered from Slack and email to watch KPIs, detect trends and anomalies, and surface what is driving a change.
  • Where it goes: after the close, Salesforce plans to bring Loops technology and talent into Tableau as part of its agentic analytics strategy.
  • Timing: expected to close in early November 2026, in Salesforce fiscal year 2027, subject to customary closing conditions.
  • Price: not disclosed. The Israeli business daily Calcalist estimated about $30 million; neither company has confirmed it (BERI).

Mark Recher, EVP and GM of Tableau, said Loops will help teams “uncover the drivers behind business performance.” Loops CEO and co-founder Tom Laufer described a shared goal of making actionable analytics available to every organization.

What Loops does, in plain terms

Most dashboards Tableau customers use today answer “what happened”. Revenue dropped 8% last week. Churn ticked up in the Midwest. Someone then has to dig through filters and drill-downs to find out why, and many teams never get past the first chart.

Loops tries to do that second step automatically. According to Salesforce Ben and TechTarget:

  • Its agent, Scout, monitors the KPIs you choose and flags unusual movement.
  • It runs root cause analysis and ranks the likely drivers by how much each one moved the metric, for example a single region, product line or channel.
  • It suggests next steps and sends alerts and short insight summaries to Slack, Microsoft Teams and email.
  • Setup is described as no-code, connecting directly to data warehouses and analytics tools. It never required Tableau.

The company was founded in 2020, is based in Foster City, California, and raised a $14 million seed round in 2022 led by Scale Venture Partners. Its website states SOC 2 Type 2 compliance and does not list public pricing.

Salesforce Loops: a team reviewing why a KPI changed on a shared screen
Explain-why agents like Loops Scout aim to answer the question teams ask after every dip: what changed, and why?

How Loops compares with what Tableau already has

Tableau customers already have some of this. The question analysts are asking is how Salesforce will stop the products from overlapping.

CapabilityWhat it does todayWhat Loops adds
Tableau PulseTracks metrics and sends insight digests by email, Slack or Teams.Ranked root cause analysis and improvement suggestions, which Pulse does not offer today (per TechTarget).
Explain DataUses statistical models to explain outliers and correlations in data points you select.Runs continuously on chosen KPIs instead of waiting for someone to click a mark.
Tableau Next and Tableau agentsSalesforce’s agentic analytics direction for Tableau, built around the Salesforce platform.A working explain-why agent and a team that has built one, which can speed up that roadmap.

Analyst David Menninger of ISG told TechTarget that Tableau needs to rationalize its KPI monitoring and agentic features once Loops is inside, or users will be confused about which tool to use. He also expects Loops to be wired into Tableau’s semantic model and single sign-on. Donald Farmer of TreeHive Strategy praised the way Loops ranks causes by impact, but said the real value depends on how quickly native integration arrives.

What we still do not know

Several practical questions about the Loops acquisition are open. Salesforce has not answered them publicly yet, so treat any confident claim you read elsewhere with care.

  • Packaging and price. Will Scout-style analysis be a standard Tableau feature, part of the premium Tableau+ tier (where Pulse’s enhanced Q&A sits today), or a paid add-on? Salesforce did not answer this when TechTarget asked.
  • Timeline. No date has been given for Loops capabilities appearing in Tableau.
  • Existing Loops customers. The announcement only talks about Tableau. It does not say whether Loops stays available as a standalone product for companies that do not use Tableau.
  • Data sources. Loops connects to warehouses directly today. Whether that stays true inside Tableau, or everything routes through Tableau or Data Cloud, is not yet clear.

BERI points out that three earlier small Salesforce AI data acquisitions, Waii, Spindle AI and Doti AI, no longer operate as standalone sites. That is not proof of what will happen to Loops, but it is a fair signal to plan for.

Why Salesforce is buying Loops, a small analytics startup

The Salesforce Loops deal fits a pattern. Salesforce has spent 2026 pushing agents into every product line: Agentforce in sales and service, the agentic integration layer it introduced at Dreamforce (first called AIforce and renamed SIforce on October 9, which we covered in our AIforce explainer), and agentic analytics in Tableau.

Analytics is a natural place for agents because the work is repetitive: watch numbers, notice changes, find the cause, tell the right person. Buying a team that already built that loop is faster than building it from scratch. It also gives Salesforce a sharper answer to competitors that sell AI-native analytics. TechTarget notes that startups such as Golden Analytics, founded by former Tableau chief product officer Francois Ajenstat, are pitching themselves to Tableau users who feel underserved.

There is also a trust angle. Tableau users have been wary of Salesforce’s stewardship since the 2019 acquisition. How well Loops is integrated, and how it is priced, will be read as a sign of how seriously Salesforce takes Tableau.

What the Salesforce Loops deal means for Tableau customers

For now, the Salesforce Loops deal changes very little day to day. Tableau keeps working as it does, and no new feature has a release date. The practical impact will come later, and it depends on something you control today: whether your metrics are worth explaining.

An agent that ranks the causes of a revenue dip is only as good as the revenue number it starts with. If finance, sales ops and marketing each calculate pipeline differently, the agent will produce confident explanations of the wrong figure. That is the same lesson many teams learned rolling out Agentforce: AI exposes messy data faster than people do.

Five things Tableau customers can do before Salesforce Loops capabilities reach Tableau:

  • Agree on KPI definitions. Write down the formula, filters and owner for the 10 to 20 metrics leadership actually watches. Put them in Tableau as governed metrics rather than in individual workbooks.
  • Fix the dimensions that explain changes. Root cause analysis splits a metric by region, product, channel, segment and rep. If those fields are blank, inconsistent or mapped differently across systems, the explanations will be noise.
  • Check where the data comes from. Many Tableau dashboards sit on CRM and ERP data. Sync errors between Salesforce and an ERP show up as “anomalies” that are really integration failures. Clean that up first.
  • Decide who receives alerts. An agent that posts to a busy Slack channel nobody owns will be muted within a month. Name an owner for each KPI and route alerts to them.
  • Ask your account team about packaging. If you are renewing Tableau in the next six months, ask directly whether Loops capabilities will be included in your edition or priced separately, and get the answer in writing.

If your data foundation already runs through Salesforce, our Salesforce Data Cloud consulting team can help unify the customer and revenue data an explain-why agent depends on.

What we see in our own Salesforce projects

In the projects our team delivers, the blocker for AI analytics is rarely the analytics tool. It is almost always the data underneath it.

For a B2B industrial supply distributor, customer-specific contract rates and volume tiers lived in dozens of spreadsheets, different reps used different versions, and pricing errors only surfaced after invoicing (read the case study). An agent asked to explain a margin dip in that business would have pointed at the wrong cause every time, because the price itself was unreliable. The fix was moving pricing into Sales Cloud and CPQ with one source of truth, connected to the ERP.

For a multi-brand retailer, marketplace orders from Amazon, Walmart and eBay were reconciled outside the CRM and support tickets sat in a separate help desk (read the case study). Revenue and service KPIs could not be explained from one place until those channels were brought into a single Salesforce org.

In both cases analytics became useful only after the data had one home. That is the order we suggest for any Tableau customer looking at what the Salesforce Loops deal could add: data first, agent second.

What existing Loops customers should do now

If you use Loops today and do not run Tableau, the Salesforce Loops announcement leaves you with open questions. Sensible steps before the early November close:

  • Check your contract for renewal dates, price protection and termination terms.
  • Make sure you can export your KPI definitions, alert rules and history.
  • Ask your Loops contact in writing whether standalone service will continue and for how long.
  • List which data sources Loops connects to, so you can plan a replacement or a move to Tableau if needed.

How this fits a wider Salesforce analytics strategy

Seen alongside its other 2026 moves, the Salesforce Loops acquisition shows Salesforce shaping a stack where data is unified in Data Cloud, actions happen through Agentforce and the agentic integration layer, and Tableau becomes the place where agents watch the numbers and explain them. Loops is a small piece, but it fills a visible gap: the “why” behind a metric.

For CRM leaders this is a reminder that analytics, data and CRM projects are converging. The KPI an agent explains in Tableau is often built on opportunity, case and order data inside Salesforce. Getting the CRM side right, from field ownership to clean integrations, is now part of getting AI analytics right. Our Salesforce implementation checklist covers the 40 decisions that most affect data quality after go-live.

What Salesforce admins and analytics teams can prepare this quarter

Admins often end up owning the data behind Tableau dashboards even when they do not build the dashboards themselves. A few practical checks will make any explain-why agent, whether it comes from the Salesforce Loops deal or elsewhere, far more useful.

  • Picklists over free text. Region, industry, product family and lead source are the fields root cause analysis splits by. If reps type them by hand, you get “TX”, “Texas” and “texas” as three different drivers. Convert the important ones to restricted picklists and clean the history.
  • Stage and close date hygiene. Pipeline metrics move when opportunities slip. If close dates are pushed in bulk at quarter end without a reason code, the agent will report the slip but cannot explain it. A simple required reason on close date changes fixes that.
  • One source for revenue. Decide whether booked revenue comes from Salesforce opportunities, CPQ quotes or the ERP invoice, and label the Tableau metric accordingly. Mixing them is the most common reason two dashboards disagree.
  • Field history on the metrics that matter. Turn on field history tracking for the handful of fields that feed leadership KPIs, so a sudden change can be traced to a person, an automation or an integration.
  • A short data dictionary. One page that says what each KPI means, where it comes from and who owns it. It costs an afternoon and saves weeks of debate once alerts start arriving in Slack.

None of this depends on when the Loops acquisition closes or how Salesforce prices it. It is the same groundwork that makes Agentforce, Tableau Pulse and ordinary reporting more reliable, which is why we usually recommend it before any AI analytics rollout.

Salesforce Loops acquisition timeline

DateEvent
August 2019Salesforce completes its acquisition of Tableau.
2020Loops is founded in Foster City, California.
2022Loops raises a $14 million seed round led by Scale Venture Partners.
September 2026Salesforce introduces its agentic integration layer at Dreamforce, first called AIforce.
October 9, 2026Marc Benioff announces that AIforce is now SIforce.
October 12, 2026Salesforce signs a definitive agreement to acquire Loops.
Early November 2026Expected close, subject to customary closing conditions. Tableau capabilities to follow.

Questions to ask your Salesforce account team

If you are among the Tableau customers watching the Loops acquisition, these are fair questions to raise at your next account review. Ask for the answers in writing, because packaging decisions are often made after a deal closes.

  • Will Loops capabilities be included in our current Tableau edition, in Tableau+, or sold as a separate add-on?
  • Will the explain-why features work on our existing published data sources, or only on data in Data Cloud or Tableau Next?
  • How will Loops alerts sit alongside Tableau Pulse, so our users are not getting two sets of notifications for the same metric?
  • What data stays inside our Salesforce or Tableau environment when the agent runs root cause analysis, and what is logged?
  • Is there an early access or pilot program we can join once the deal closes?

None of these questions commit you to anything. They do make sure the Salesforce Loops features are evaluated on your terms rather than discovered in a renewal quote.

Frequently asked questions

What is Loops, the company Salesforce is acquiring?

Loops is an AI-powered analytics platform founded in 2020 and based in Foster City, California. Its agent, Scout, monitors chosen KPIs, explains why they change by ranking the likely causes, and sends alerts and insight summaries to Slack, Microsoft Teams and email.

When will the Salesforce Loops acquisition close?

Salesforce expects the deal to close in early November 2026, in its fiscal year 2027, subject to customary closing conditions.

How much is Salesforce paying for Loops?

The price was not disclosed. One estimate reported by the Israeli business daily Calcalist put it at about $30 million, but neither company has confirmed that figure.

Will Loops features be included in Tableau?

Salesforce plans to bring Loops technology and talent into Tableau after the close. It has not said when the features will appear or whether they will be standard, part of a premium tier or a paid add-on.

How is Loops different from Tableau Pulse?

Tableau Pulse tracks metrics and sends insight digests. Loops adds continuous root cause analysis that ranks the drivers behind a metric change and suggests next steps, which Pulse does not do today.

What should Tableau customers do before Loops arrives?

Agree on KPI definitions and owners, clean up the dimensions used to explain changes, fix CRM and ERP sync problems that create false anomalies, and ask your account team how Loops capabilities will be packaged before your next renewal.

Getting started

Ashapura Softech works with US businesses from Irving, Texas, with a development center in Ahmedabad. If you want your Salesforce data ready for agentic analytics, from clean integrations to unified customer data, our Salesforce CRM development and integration services are a good place to start.

About the author. Ankur Pandey is Marketing Manager at Ashapura Softech, a Salesforce and Odoo consulting partner based in Irving, Texas, with a delivery center in Ahmedabad, India. He works with the company’s Salesforce delivery teams on CRM, Data Cloud and integration projects for US businesses and writes about what Salesforce product changes mean in practice.

Disclosure. Ashapura Softech is a Salesforce consulting partner. We have no business relationship with Loops and were not paid to write this article. Facts were checked against the sources below on October 19, 2026, and the article will be updated when the deal closes or Salesforce announces packaging.

Sources

Categories
Salesforce

Salesforce Phishing-Resistant MFA: The Complete 2026 Guide for Admins

Last updated: 16 October 2026. Checked against Salesforce Help articles 005321563, 005388907 and 005390721 on the same date.

Key takeaways

  • Salesforce now requires phishing-resistant MFA for privileged users: anyone with the System Administrator profile, or the Modify All Data, View All Data, Customize Application or Author Apex permission.
  • Only passkeys, built-in authenticators (Touch ID, Face ID, Windows Hello), FIDO2 security keys, certificate-based authentication and admin-generated temporary codes count. Salesforce Authenticator push, TOTP apps and SMS do not.
  • Production enforcement rolled out by release group between 20 July and 3 September 2026. In October, Salesforce said there are “no indefinite delays” and is only working with a small group of customers on extensions.
  • SSO users are covered only if the identity provider sends a phishing-resistant AMR or ACR value such as fido2, hwk or phr.

Most Salesforce admins had MFA switched on years ago, so the 2026 change looked small on paper. It wasn’t. The new rule targets the people who can change everything in an org, and it throws out the methods most of them were using. An admin who approved a Salesforce Authenticator push every morning in spring now gets a “Create a Passkey” screen with no other option.

This guide explains the Salesforce phishing-resistant MFA requirement in full: who is in scope, the official dates, which methods pass, how SSO is judged, what breaks for integrations, sandboxes and mobile, and how to fix the login problems admins are reporting. Every rule below is taken from Salesforce’s own Help documentation, linked in the sources at the end.

What Salesforce phishing-resistant MFA means, in plain terms

Normal MFA asks for a second proof after your password, such as a six-digit code or a tap on your phone. The weakness is that a fake login page can sit between you and Salesforce. You type your password and code into the fake page, and the attacker replays them to the real site within seconds. Push approvals have the same problem, because people approve prompts they did not start.

Phishing-resistant methods are built on FIDO2 and WebAuthn. When you register a passkey or security key, it is tied to the real Salesforce domain. If a lookalike site asks for it, the browser simply won’t hand it over. There is nothing for the user to type, so there is nothing for an attacker to steal and replay.

How Salesforce phishing-resistant MFA stops a fake login page compared with codes and push approvals
Codes and push approvals can be relayed by a fake page. Salesforce phishing-resistant MFA methods cannot.

That is why Salesforce phishing-resistant MFA is aimed at privileged users first. One stolen admin session can export every record or change security settings for the whole org.

The Salesforce MFA requirement in 2026: two rules, not one

Salesforce enforces two separate controls, and they are easy to mix up.

RuleWho it coversWhat satisfies it
MFA for all employee UI loginsEvery internal user who logs in through the browser or mobile appAny standard MFA method, such as Salesforce Authenticator, a TOTP app, a passkey or a security key
Phishing-resistant MFA for privileged usersAdmins and users with any of the four high-risk permissionsPasskey, built-in authenticator, FIDO2 security key, certificate-based authentication or an admin-generated temporary code

Once enforcement reaches your org, the setting “Require multi-factor authentication (MFA) for all direct UI logins to your Salesforce org” is switched on by Salesforce. Without an approved extension, admins can’t turn it off. It shows as greyed out or hidden.

Salesforce MFA enforcement dates

According to Salesforce Help article 005321563, Salesforce phishing-resistant MFA for privileged users rolled out in this order:

Release groupEnforcement date
Preview sandboxes10 July 2026
R020 July 2026
R127 to 28 July 2026
R2a29 to 30 July 2026
R2b3 August 2026
Japan and Korea1 September 2026
All other orgs3 September 2026

Two details matter here. First, production was originally planned for 1 July and moved to 20 July after a bug that asked users with existing security keys to register new ones. Second, trial orgs converted to a paid subscription do not get the usual 30-day grace period. If you converted a trial recently, assume enforcement is already live.

Where things stand in October 2026

On 15 October, Salesforce Ben reported that some customers had received notices suggesting enforcement could slip further. Salesforce pushed back on that. A spokesperson said “there are no indefinite delays on anything” and that all controls will be implemented. Salesforce says it is working with a small number of customers who already had extensions over the next six to eight months, and that more than 90% of customers have adopted MFA.

The practical reading: if you have not received an extension from Salesforce in writing, treat the rule as active in your org today.

Who must use Salesforce phishing-resistant MFA

Scope follows permissions, not job titles. A finance analyst with View All Data is a privileged user in Salesforce’s eyes, even if nobody on your team calls them an admin.

Flowchart showing who needs Salesforce phishing-resistant MFA based on admin profile and permissions
Salesforce phishing-resistant MFA applies to anyone with the System Administrator profile or one of four high-risk permissions.
If the user has…Requirement
System Administrator profile, or a clone of itPhishing-resistant MFA
Modify All DataPhishing-resistant MFA
View All DataPhishing-resistant MFA
Customize ApplicationPhishing-resistant MFA
Author ApexPhishing-resistant MFA
None of the above, logs in through the UIStandard MFA

The permission can come from a profile, a permission set or a permission set group. The rule covers direct UI logins and SSO logins, in production and in sandboxes. It also covers Chatter Plus users and internal users who reach an Employee Community or Experience Cloud site.

Not in scope:

  • External users on Experience Cloud, Chatter External and Chatter Free licenses
  • Developer Edition, trial, scratch and other non-paid orgs
  • Pure API logins that never touch the login page (see the integrations section below)

How to find every privileged user in your org

Start with a list, not assumptions. This SOQL query returns every active user who falls under Salesforce phishing-resistant MFA because they hold one of the four permissions. Profiles are included too, because every profile has an underlying permission set that shows up in PermissionSetAssignment.

SELECT Id, Name, Username, Profile.Name
FROM User
WHERE IsActive = true
AND Id IN (
  SELECT AssigneeId
  FROM PermissionSetAssignment
  WHERE PermissionSet.PermissionsModifyAllData = true
  OR PermissionSet.PermissionsViewAllData = true
  OR PermissionSet.PermissionsCustomizeApplication = true
  OR PermissionSet.PermissionsAuthorApex = true
)

Run it in Developer Console or Workbench, then export the result. If you prefer clicks over code, the User Access and Permissions Assistant from AppExchange gives the same answer. Also pull a list of anyone holding “Waive Multi-Factor Authentication for Exempt Users”, because that permission no longer protects them.

Before you buy a single security key, look at the list and ask who really needs that access. Removing View All Data from five people who only needed it once is the cheapest compliance step you will take.

Which methods count as Salesforce phishing-resistant MFA

MethodStandard usersPrivileged users
Passkey, device-bound or synced through a FIDO2-compliant password managerYesYes
Built-in authenticator (Touch ID, Face ID, Windows Hello)YesYes
Security key (FIDO2 or WebAuthn, such as a YubiKey)YesYes
Certificate-based authentication (x.509)YesYes
Admin-generated temporary verification codeYesYes, for short-term recovery
Salesforce Authenticator pushYesNo
TOTP apps (Google Authenticator, Microsoft Authenticator, Authy)YesNo
SMS, email or phone codesNo (weak)No

Passkey, built-in authenticator or security key: which to choose

OptionWorks best forWatch out for
Device-bound passkey or built-in authenticatorAdmins who work from one laptop most of the timeIt stays on that device. A new laptop means registering again.
Synced passkey in a password manager (Apple Passwords, Google Password Manager, 1Password, Bitwarden)Admins who move between several machinesThe password manager account itself now protects your Salesforce org, so lock it down too.
Hardware security keyShared desks, locked-down machines, consultants working across many orgsKeys get lost. Buy two per person and register both.

Our general advice: give every privileged user two methods. A built-in authenticator for daily use plus a hardware key in a drawer covers almost every lockout story.

How to set up Salesforce phishing-resistant MFA

Step 1: Admin turns on the methods (one time)

  1. In Setup, type Identity Verification in Quick Find.
  2. Enable built-in authenticators and security keys. These settings apply to all users. You can’t limit them to admins.
  3. Optional but recommended: enable Allow passwordless login with passkeys. If you run Experience Cloud sites, read the external user note in the troubleshooting section first.
  4. Save.

Step 2: Each privileged user registers a method

  1. Open your personal Settings and go to Advanced User Details.
  2. For a passkey or built-in authenticator, find Built-in Authenticators, click Add, then Create Passkey. Confirm with your fingerprint, face or device PIN and give it a clear name, such as “Work MacBook”.
  3. For a hardware key, find Security Key (U2F or WebAuthn), click Register and tap the key.
  4. Repeat for a second method.

Users who already have an MFA method must verify with it before adding a new one. Users with nothing registered get a one-time code by SMS or email to get started.

Step 3: Test before you rely on it

Log out, log back in with the new method, and try it from a second browser. Do this in a sandbox first if your team has never used passkeys.

How SSO logins are judged: AMR and ACR values

SSO does not exempt anyone from Salesforce phishing-resistant MFA. When users sign in through Okta, Microsoft Entra ID, Ping or ADFS, Salesforce reads the authentication signals the identity provider sends in the SAML response or OIDC ID token. These are the AMR (Authentication Method Reference) and ACR (Authentication Context Class Reference) values.

Signal levelAccepted AMR values
Phishing-resistant (needed for privileged users)cert, face, fido, fido2, fpt, hwk, iris, passkey, phr, pki, pop, pwlesspasskey, retina, sc, smartcard, smartcardpki, softwarepki, swk, tlsclient, x509
Standard (enough for other users)mfa, mobiletwofactorcontract, okta_verify, pin, pgp, publickey, rsa, timesynctoken, user, vbm
Weak (not accepted)pwd, sms, tel, email

The accepted ACR list for phishing-resistant MFA is almost the same: fido, fido2, fpt, hwk, passkey, phr, pki, pwlesspasskey, retina, smartcard, smartcardpki, softwarepki, swk, tlsclient and x509.

Formatting rules that trip people up:

  • SAML AMR: the attribute name must contain “amr” or “authnmethodsreferences”. Separate multiple values with semicolons. Comma-separated values are not parsed.
  • OIDC AMR: send an array, for example ["hwk","mfa"]. Matching is exact.
  • ACR: matching is case-insensitive and uses a “contains” check, so a full URN that includes the value still works.

To check what your identity provider actually sends, open Login History and look at the AMR and ACR fields for a recent SSO login, or paste a response into the SAML Assertion Validator. If the right value isn’t there yet, you can turn on Use Salesforce MFA for this SSO Provider as a stopgap. Users then complete Salesforce MFA on top of the identity provider until the signal is fixed.

Extensions and the end of the Waive MFA permission

The “Waive Multi-Factor Authentication for Exempt Users” permission used to switch MFA off for specific accounts. It no longer does. Exceptions, such as automated UI testing tools, now need approval from Salesforce Support.

There is a developer side effect too. The API field PermissionsBypassMFAForUiLogins is removed unless your org has an extension, so Apex, metadata or scripts that reference it can fail to compile. Search your codebase and deployment packages for it.

Salesforce has granted two kinds of temporary extension:

  • MFA For All Employees extension: covers non-privileged users only.
  • Phishing-Resistant MFA (Passkeys) extension: temporarily lets privileged users use standard MFA instead.

Even with an extension, users may still see passkey prompts in some combinations of settings. Salesforce Help article 005388907 lists six scenarios and says this is expected behaviour, not a bug.

Integration users and API connections

Salesforce phishing-resistant MFA, like standard MFA, applies to UI logins. A connection that never shows a login page is not prompted. The catch is that some OAuth flows do involve an interactive login.

Connection typeAffected by MFA?
OAuth JWT bearer flowNo
OAuth client credentials flowNo
OAuth web server and user-agent flowsYes, they use a UI login
Username-password flow and SOAP login() callsListed by Salesforce as affected in its troubleshooting article

If an integration stopped working after enforcement, open Login History and check the Login Subtype and API Type fields for that user. The long-term fix is to move the integration to the JWT bearer or client credentials flow and keep the integration user API-only. This matters most for older middleware, ETL tools and custom scripts. If you are reviewing those connections anyway, our guide to Salesforce integration with other tools covers the main patterns.

Sandboxes and mobile: two places admins get surprised

Sandbox refreshes

Registered MFA methods do not copy into a refreshed sandbox. SAML SSO settings are partly copied but disabled, and the Recipient URL is updated. After every refresh, admins need to log in directly, register a passkey or key, then switch SSO back on. Since a late-July patch, a refreshed sandbox does inherit any extension your production org has, so you don’t need to request it twice.

Mobile apps and the Mobile SDK

Since July 2026, the mobile login shows the username first and then asks for a passkey or password. This change can break automated UI tests that script the old login screen. If you build on the Salesforce Mobile SDK, version 13.2.0 and earlier block admins in the default WebView unless advanced authentication is set up in My Domain. Version 13.2.1 adds a “Login for Admin” option.

Troubleshooting Salesforce passkey problems

What the user seesLikely causeFix
“Create a Passkey” with no option to choose another methodThe user is privilegedExpected. Register a passkey or security key.
“Your account requires a passkey for enhanced security” after SSOIdentity provider did not send a phishing-resistant AMR or ACR valueFix the identity provider signal, or use Salesforce MFA for that SSO provider for now
Passkey works on the laptop but not on a second deviceDevice-bound passkey, not syncedUse a synced passkey, a hardware key, or log in by scanning a QR code with your phone
Browser says “This device cannot be used”, or nothing happens after Create PasskeyUnsupported browser, no screen lock or PIN on the device, or passkeys disabled in the OSUse current Chrome, Edge, Firefox or Safari, set a device PIN, or register from a second device via QR code
Passkey prompt after a sandbox refreshVerifiers don’t carry overRegister again and re-enable SSO
External community users see a passkey promptPasswordless login with passkeys is onTurn off “Allow passwordless login with passkeys” or confirm the user’s license type
User with the Waive MFA permission is asked for a passkeyThat permission no longer exempts anyoneRegister a method, or request an exception from Salesforce Support
Admin is locked outLost device or keyAnother admin issues a temporary verification code. If you are the only admin, contact Salesforce Support with your org ID and username.

To replace a broken passkey, log in with another method, open your personal settings, delete the old passkey under Built-in Authenticators (or on the Passkeys page), then add a new one.

A 30-day Salesforce phishing-resistant MFA rollout plan

If your org is still catching up, this four-week plan gets every privileged user compliant without surprise lockouts.

30-day Salesforce phishing-resistant MFA rollout plan for admins, week by week
A four-week Salesforce phishing-resistant MFA rollout plan.
  1. Week 1, inventory: run the SOQL query, list users with the Waive permission, list SSO providers and integration users. Remove access that nobody needs.
  2. Week 2, enable and pilot: turn on built-in authenticators and security keys in a sandbox. Have two admins register two methods each and run their normal work for a few days.
  3. Week 3, fix SSO and integrations: confirm your identity provider sends a phishing-resistant value for admins. Move any integration on a UI-based OAuth flow to JWT bearer or client credentials. Search code for PermissionsBypassMFAForUiLogins.
  4. Week 4, roll out and document: register every privileged user in production, give your help desk the “Manage Multi-Factor Authentication in User Interface” permission so they can issue temporary codes, and write down who to call if the last admin gets locked out.

If MFA is one of many admin and security jobs piling up, our Salesforce managed services team can run the privileged access review and rollout with you. Teams planning a new org can build these controls in from day one using our Salesforce implementation checklist.

Salesforce phishing-resistant MFA FAQs

Does Salesforce Authenticator count as phishing-resistant MFA?

No. Salesforce Authenticator push approvals and codes still count as standard MFA for regular users. Privileged users need a passkey, built-in authenticator, security key or certificate-based authentication.

Who counts as a privileged user in Salesforce?

Anyone with the System Administrator profile or a clone of it, or the Modify All Data, View All Data, Customize Application or Author Apex permission, whether it comes from a profile, a permission set or a permission set group.

When was Salesforce phishing-resistant MFA enforced?

Preview sandboxes on 10 July 2026, then production by release group from 20 July to 3 September 2026. Japan and Korea followed on 1 September. Salesforce says it is working with a small number of customers on extensions over the six to eight months following October 2026.

Can we still exempt users from MFA?

The Waive Multi-Factor Authentication for Exempt Users permission no longer works as an exemption. Exceptions, such as for automated testing tools, must be approved by Salesforce Support.

Does SSO satisfy the requirement?

Only if your identity provider sends an accepted AMR or ACR value. For privileged users, that value must be phishing-resistant, such as fido2, hwk, passkey or phr. A generic “mfa” value only covers standard users.

Do integration users need phishing-resistant MFA?

API logins are not prompted. Integrations using the JWT bearer or client credentials flow are unaffected. Integrations using OAuth web server or user-agent flows go through a UI login and can be blocked, so move them to an API-only flow.

Why does Salesforce keep asking me to create a passkey?

Either you are a privileged user, your SSO login did not send a phishing-resistant signal, or you are in a refreshed sandbox where your methods did not carry over. The troubleshooting table above maps each message to a fix.

What happens if a privileged user loses their only passkey?

Another admin can generate a temporary verification code so they can log in and register a new method. If there is no other admin, contact Salesforce Support with your org ID and admin username.

Sources

Categories
Agentforce Salesforce

Salesforce Koa: What the CRM Reasoning Model Means for Your Org

Last updated: 15 October 2026

Short answer

Salesforce Koa is Salesforce’s first CRM reasoning model. It is NVIDIA Nemotron 3 Super, post-trained by Salesforce on synthetic CRM scenarios, and it runs inside Salesforce’s own trust boundary. Koa is built to help Agentforce agents pick the right action on multistep work like qualifying a lead or routing a case. Today it is limited to pilot customers. General availability is expected in winter 2026 in US regions, and pricing has not been announced.

Salesforce announced a lot at Dreamforce ’26, and Koa got less attention than Claudeforce or Agentforce Coworker. For admins and architects, though, it may be the announcement with the most day-to-day impact. Koa changes a question most teams never had to ask: which model should sit behind each of our agents?

This guide covers what Salesforce Koa is, how it was built, what Salesforce claims it does better, the three places it will show up in your org, and a practical checklist for getting ready before general availability. Where Salesforce has not published something yet, such as pricing, we say so instead of guessing.

What is Salesforce Koa?

Koa is a large language model that Salesforce trained specifically for CRM work. Salesforce describes it as a “CRM reasoning model” for Agentforce agents. Its job is not to write marketing copy or summarize long documents. It is meant to reason through tasks that involve several steps and several tool calls, such as updating an opportunity, routing a case to the right queue or scheduling a follow-up after a call.

Salesforce first showed Koa on 15 September 2026 at Dreamforce, as part of a wider announcement with NVIDIA. On 13 October 2026, Salesforce’s Chief Data Officer Michael Andrew gave more detail on how it was trained in a Salesforce newsroom interview.

Salesforce Koa at a glance

ItemWhat we know (October 2026)
Announced15 September 2026, Dreamforce ’26
Base modelNVIDIA Nemotron 3 Super (120B total parameters, 12B active, up to 1M token context)
Training dataSynthetic scenarios modeled on Salesforce CRM deployments across 14+ industries. Salesforce says no customer data was used.
Where it runsSalesforce infrastructure. Salesforce controls the model weights.
StatusPilot customers only, inside Agentforce
General availabilityExpected winter 2026, US regions, with an open beta planned shortly after
Pilot customers named1-800Accountant, Baxter Credit Union, Engine, Formula 1, UChicago Medicine, Xero
PricingNot announced

How Salesforce Koa was built

Salesforce did not train Koa from scratch. It started with NVIDIA Nemotron 3 Super, an open model NVIDIA released in March 2026. Nemotron 3 Super is a hybrid design that mixes Mamba-2 layers, mixture-of-experts layers and a small number of attention layers. Only about 12 billion of its 120 billion parameters are active for any one request, which keeps it cheaper to run than a dense model of the same size.

Salesforce then post-trained it in two stages:

  • Supervised fine-tuning. The model learned from worked examples of CRM tasks done correctly.
  • Reinforcement learning with GRPO. The model practised tasks in simulated environments and was rewarded for choosing the right tools and actions. Salesforce used NVIDIA’s NeMo RL, NeMo Gym and NeMo AutoModel tooling for this.

The training scenarios are synthetic. Salesforce built them from what it has learned running CRM deployments for close to three decades, across manufacturing, financial services, healthcare, travel and other industries. That matters for two reasons. First, it is why Salesforce can say no customer data went into training. Second, it is the main open question: synthetic scenarios are only as good as how closely they match messy real orgs, with their custom objects, half-used fields and years of validation rules.

What Salesforce Koa does better, according to Salesforce

Salesforce tested Koa on its own CRM benchmark, which it calls CRM Bench. The tasks are everyday CRM actions: updating an opportunity, routing a case, scheduling a follow-up. Here is what Salesforce reports.

Salesforce Koa CRM Bench claims: 3x fewer errors, 2.1x better context recall, 15% better long-conversation context, 11% more precise action calling
Salesforce’s published Koa figures. These are vendor numbers, not independent tests.
  • 3x fewer errors on CRM actions, while matching or beating leading models
  • 11% more precise at calling the right action, compared with default general-purpose models
  • 2.1x more reliable at recalling customer context
  • 15% better at keeping context across long back-and-forth conversations

Treat these as a reason to test Salesforce Koa, not as proof. Salesforce has not published the full benchmark, named the models it compared against for the “3x” figure, or shared independent results. Koa also runs at temperature 0, which means it gives the same answer to the same input. That is good for audit and repeatability, and it is one reason a model can look more reliable on structured tasks.

What we find most useful in these numbers is where the gains are. They are all about choosing actions and holding onto context. That lines up with the failures we see most often when teams pilot agents: the agent picks the wrong flow, updates the wrong record, or forgets what the customer said three turns earlier.

How to access Salesforce Koa in your org

According to the Koa product page, Salesforce Koa will be available in three places. All three are opt-in.

Three ways to access Salesforce Koa: Data Cloud Generative Models catalogue, org-wide model provider in Setup, and per agent in Agentforce Builder
The three access points Salesforce lists for Salesforce Koa.
  1. Data Cloud Generative Models catalogue. Koa is listed as a managed model you can choose for prompt templates, AI apps and agents.
  2. Org-wide, in Setup. Koa appears as a fourth model provider. Turning it on here affects every agent in the org.
  3. Per agent, in Agentforce Builder. You can choose Koa for a single agent, sub-agent or agent router.

Our advice is to start with option 3. Switching the whole org to a new model in one step makes it hard to tell which change caused a problem. Running Salesforce Koa on one agent, next to an unchanged copy on your current model, gives you a clean comparison.

One note of caution: the product page also describes Koa as available “out of the box” in the Data Cloud catalogue, while the same page and the press release say it is pilot-only until general availability. If you do not see Koa in your org yet, that is expected. Check with your Salesforce account team if you want to join the pilot.

Salesforce Koa vs Claudeforce vs your current Agentforce model

Salesforce Koa arrives at the same time as Claudeforce, and the two are easy to confuse. They solve different problems.

Salesforce KoaClaudeforceDefault Agentforce models
What it isSalesforce’s own CRM reasoning modelSalesforce and Anthropic partnership: Salesforce in Claude, plus Claude inside SalesforceGeneral-purpose models from partner providers
Best atPicking the right CRM action on multistep tasksWorking with Salesforce data from Claude, broad reasoning and writingBroad tasks: summaries, drafting, Q&A
Where it runsSalesforce infrastructure, Salesforce-controlled weightsAnthropic models through the Salesforce trust layerProvider models through the trust layer
StatusPilot, GA expected winter 2026 (US)Salesforce in Claude in betaGenerally available
PricingNot announcedDepends on Claude plan and Salesforce usageThrough your Agentforce and Flex Credits agreement

In practice, many orgs will end up using more than one. A service agent that routes and updates cases might run on Koa, while a sales rep drafts emails and reviews pipeline in Claude. Our Claudeforce vs Agentforce comparison explains how those two layers fit together.

Should you switch your agents to Salesforce Koa?

Not yet, for most orgs, simply because you cannot until general availability. The better question is which agents are good candidates once you can. Based on what Salesforce has published, Salesforce Koa looks like a strong fit when:

  • The agent’s main job is taking actions in Salesforce, not writing long text
  • Each task needs two or more tool calls in the right order, such as look up the account, check entitlements, then route the case
  • Wrong actions are costly, for example a case sent to the wrong queue or an opportunity stage changed by mistake
  • Conversations run long and the agent keeps losing earlier details
  • Your security team prefers a model whose weights stay inside Salesforce

It is a weaker fit, or at least unproven, for open-ended writing, analytics questions across large datasets, and work outside the CRM. For those, a frontier model such as Claude may still be the better choice.

How to get your org ready for Salesforce Koa: a 6-step checklist

None of these steps depends on Koa pricing or release dates, so you can start now. They also improve your current agents, whichever model you use.

  1. List every agent and the model behind it. Include Agentforce agents, prompt templates, any Einstein features and any Claude or OpenAI connections. Many teams have never written this down.
  2. Pick two or three action-heavy workflows. Lead qualification, case routing and opportunity updates are the obvious ones. These are where Koa’s claimed gains apply.
  3. Build a test set from real cases. Collect 30 to 50 real examples per workflow, with the correct outcome for each. Strip out personal data. This is the only way to check Salesforce’s benchmark against your own org.
  4. Tighten permissions before you add any model. Review the permission sets, sharing rules and field-level security your agents run under. A more capable model with over-broad access is a bigger risk, not a smaller one. Our Agentforce readiness checklist covers this in detail.
  5. Set your pass and fail rules. Decide in advance what error rate is acceptable, when the agent must hand over to a person and what you need logged for audit.
  6. Plan for cost uncertainty. Salesforce has not said whether Koa will cost more or less than current models. If you are on Flex Credits, track credit use per agent now so you can compare once Koa pricing is public. Agent calls from outside Salesforce are also getting a new meter, which we explain in our guide to AIforce (now SIForce) and your implementation.

Salesforce Koa timeline

  • 11 March 2026: NVIDIA releases Nemotron 3 Super, the base model for Koa
  • 15 September 2026: Salesforce announces Koa at Dreamforce ’26
  • October 2026: Post-trained NVIDIA models reach select Missionforce customers. Salesforce shares more training detail on 13 October.
  • Now: Koa available to select pilot customers in Agentforce
  • Winter 2026: General availability expected in US regions, followed by an open beta

Planning an Agentforce rollout? We help US teams pick the right workflows, set up permissions and test agents before they go live. Book a free 30-minute consultation and we will review your current agents with you.

Related guides

Frequently asked questions

What is Salesforce Koa?

Koa is Salesforce’s first CRM reasoning model. It is NVIDIA Nemotron 3 Super, post-trained by Salesforce on synthetic CRM scenarios, and it is designed to help Agentforce agents choose the right actions on multistep CRM tasks.

When will Salesforce Koa be generally available?

Salesforce expects general availability in winter 2026 in US regions, with an open beta to follow. Until then it is limited to select pilot customers.

Is Koa trained on customer data?

No, according to Salesforce. Koa was trained on synthetic scenarios built from Salesforce’s experience with CRM deployments, and Salesforce says no customer data crosses its trust boundary during training or inference.

How much does Salesforce Koa cost?

Salesforce has not announced Koa pricing. If you use Flex Credits today, track credit use per agent so you can compare costs once pricing is published.

Is Salesforce Koa the same as Claudeforce?

No. Koa is Salesforce’s own model, running on Salesforce infrastructure. Claudeforce is Salesforce’s partnership with Anthropic that brings Salesforce data into Claude and Claude into Salesforce. Many orgs may use both for different jobs.

How do I turn on Koa in Agentforce?

Once it is available to your org, you will be able to choose Koa per agent in Agentforce Builder, set it as an org-wide model provider in Setup, or pick it from the Data Cloud Generative Models catalogue. We recommend testing on one agent first.

The bottom line

Salesforce Koa is the company’s bet that a smaller model trained on CRM work can beat general models at the jobs agents actually do inside Salesforce. The claims are promising but come only from Salesforce so far, and there is no price yet. The smart move now is not to wait. Map your agents, pick the action-heavy workflows, build a real test set and fix permissions. When Salesforce Koa reaches general availability, you will be able to test it in days instead of months.

How we put this guide together

Product facts in this guide come from Salesforce’s 15 September 2026 Koa press release and its Koa product page, with base model details from NVIDIA’s Nemotron 3 Super model card. Benchmark figures are Salesforce’s own and are labeled that way. Pricing, Koa’s exact place in Agentforce packaging and independent test results were not public at the time of writing, so we have not guessed at them. We will update this page when Koa reaches general availability.

Sources:

About the author

Ankur Pandey is Marketing Manager at Ashapura Softech, a certified Salesforce and Odoo partner with offices in Irving, Texas and Ahmedabad, India. He writes about Salesforce, Zoho and ERP planning for US mid-market businesses.

Categories
Salesforce

Salesforce Outage September 2026: What Happened and What Admins Should Check Now

Last updated: 14 October 2026

Short answer

The Salesforce outage on 16 September 2026 started at 07:50 UTC (12:50 AM Pacific), on day two of Dreamforce, and disrupted logins, APIs and support case creation on hundreds of instances across all regions. Service was healthy again from 15:26 UTC, about seven and a half hours later, and Salesforce marked the incident resolved at 18:59 UTC. Salesforce has said requests stalled while waiting on an internal login service, but as of 14 October it has not published a full root cause. If you run Salesforce, the job now is to check that scheduled jobs, integrations and flows that failed that day were picked up again.

Most Salesforce teams found out about this one the hard way. Integration alerts started firing overnight in the US, users in Europe and Asia could not log in, and the usual escape route, opening a case with Salesforce Support, was broken too. By the time the US woke up, the fix was already rolling out, so plenty of people only saw the tail end and assumed it was over. Integrations that failed loudly that night are worth a second look, and error handling is the first thing we review on any Salesforce CRM integration.

It wasn’t, at least not for every org. Salesforce itself reported that scheduled jobs were not running as expected for some customers even after access came back. That is the part of a Salesforce outage that quietly causes problems weeks later: a nightly sync that never ran, a renewal reminder that never went out, a batch job that stopped halfway.

This guide covers what happened, what is confirmed and what is not, a practical checklist for admins, how to check whether Salesforce is down right now, and what to put in place before the next one.

What happened during the Salesforce outage on 16 September 2026

Salesforce tracked the incident on its Trust status site as incident 20004433. The timeline below is built from Salesforce’s status updates as reported by Salesforce Ben and CIO.

Time (UTC)Time (Pacific)What Salesforce reported
07:5012:50 AMDisruption begins. Logins, APIs and support case creation affected on multiple instances in all regions.
09:102:10 AMRequests are stalling while waiting on an internal login service, using up server resources. A rolling restart is tried on one instance.
09:572:57 AMSalesforce suspects an external dependency failure affecting its legacy login server and blocks an API endpoint. Its third-party infrastructure provider reports no issues.
10:183:18 AMA core system component is under increased load, limiting how many requests it can process.
10:563:56 AMA fix is validated on a test instance.
11:194:19 AMFleetwide rollout of the fix begins, with a permanent code-level fix in parallel.
11:414:41 AMRollout continues region by region. GovCloud is confirmed as not affected.
14:007:00 AMSome instances need manual restarts. Scheduled jobs are not running as expected for some customers.
15:268:26 AMService healthy from this point, as Salesforce later confirmed.
15:398:39 AMRemaining impact narrowed to a subset of Hyperforce instances.
18:5911:59 AMIncident marked resolved.

Salesforce outage September 2026 timeline: disruption from 07:50 UTC, service healthy from 15:26 UTC, resolved at 18:59 UTC

So the hard outage lasted about 7 hours 36 minutes, and the incident stayed open for roughly 11 hours while Salesforce confirmed every instance was stable.

How big was the outage?

Salesforce’s Trust site described “hundreds of instances” affected across all regions. Reports that track the Trust record say the list peaked above 1,000 instances during the incident, and was later trimmed to 912 of about 3,243 once Salesforce removed instances that had never been affected. Salesforce also said the impact was “narrower than initially understood”.

A few details matter for working out whether your org was hit:

  • GovCloud was not affected, according to Salesforce’s updates.
  • Hyperforce instances recovered last. The final stretch of the incident involved a subset of Hyperforce instances.
  • It was an application problem, not a network one. Independent monitoring cited by CX Today saw HTTP 503 errors and timeouts with no network degradation.
  • Well-known companies felt it. CX Today reported effects at PayPay, Sage (UK and Ireland support teams), Vonage and Veeva CRM users. In one case, web forms kept collecting leads but the data had to sync into Salesforce after recovery.

What caused the Salesforce outage?

Here is what Salesforce has said, and nothing more. Requests stalled while waiting for a response from an internal login service, and those stalled requests used up server resources. Early on, Salesforce suspected an external dependency affecting its legacy login server, but its third-party infrastructure provider reported no issues. Salesforce also pointed to increased load on a core system component.

Salesforce apologised and said it would establish “the technical trigger, the underlying cause, and preventive action”. As of 14 October 2026, we could not find a published root cause analysis on the Trust site or in Salesforce’s public statements. Until one appears, treat any detailed explanation you read online as a guess.

If your business depends on Salesforce, ask your account executive for the post-incident report. It is the only way to know whether the fix covers the kind of failure that hit your org.

A second incident a few days later: Data Cloud

Around 19 September, Data Cloud activations in the US-East-1 region were down for 13 hours 41 minutes. Salesforce said that problem came from a third-party cloud storage service and that it was preparing an alternate storage option. If you use Data Cloud (now Data 360) activations to feed marketing or ads, check that segments activated that week actually reached their destinations.

Not related: Chrome and Edge freezing

Around the same time, some users on Chrome and Edge 153 saw the Salesforce interface freeze. Salesforce linked that to an external Chromium update, separate from the outage. A Chrome fix was released first. If users still report freezing, check their browser version before assuming it is another outage.

Post-outage checklist for Salesforce admins

If your org was on an affected instance, logins working again does not mean everything caught up. Here is the order we would work through it. Most of it takes an afternoon.

1. Confirm your instance and the impact window

Find your instance on status.salesforce.com and note when it was affected. Use that window for every check below. Write it down; you will need it for your incident log and for any conversation with Salesforce.

2. Check scheduled Apex and scheduled jobs

Salesforce reported scheduled jobs not running as expected after access returned. Run this query in the Developer Console or Workbench and look for jobs whose last run is before the outage when they should have run during or after it:

SELECT CronJobDetail.Name, CronJobDetail.JobType, State, PreviousFireTime, NextFireTime
FROM CronTrigger
ORDER BY PreviousFireTime

Any job that missed its slot needs a manual run. Also check Setup, then Scheduled Jobs, for dashboard refreshes and reporting snapshots, which fail quietly.

3. Check batch and queueable Apex that failed or stopped

SELECT ApexClass.Name, JobType, Status, ExtendedStatus, NumberOfErrors, CreatedDate
FROM AsyncApexJob
WHERE CreatedDate >= 2026-09-16T07:00:00Z
AND CreatedDate <= 2026-09-17T00:00:00Z
AND Status IN ('Failed', 'Aborted')

Batch jobs that failed partway may have updated some records and not others. Re-run them only if the job is safe to run twice; otherwise fix the gap by hand or with a targeted job.

4. Review paused and failed flows

Go to Setup, then Paused and Failed Flow Interviews. Scheduled-triggered flows that did not run will not show up here, so compare them against what should have happened that day (for example, reminder emails or status updates).

5. Reconcile integrations

This is where most of the damage hides. For each integration (ERP, billing, marketing automation, support tools):

  • Check the integration’s own error log for failed calls in the outage window.
  • Look for duplicates. Retry logic often creates the same record twice when the first call actually succeeded.
  • Look for gaps. Records created in the other system during the outage may never have reached Salesforce.
  • Check the integration user’s Login History (Setup, then Login History) for failures, and make sure no one rotated credentials in a panic during the outage.

6. Check web-to-lead, web-to-case and inbound email

Forms and email-to-case are the front door for customers. Compare the number of leads and cases created that day against a normal day. If your form tool stores submissions, re-send the missing ones.

7. Confirm support cases you tried to log

Case creation on Salesforce Help was disrupted. If you opened a case with Salesforce during the outage, confirm it actually exists.

8. Write a short incident log

Record the impact window, what failed, what you fixed and what is still open. It takes 20 minutes and makes the next outage much easier, and it is the evidence you need if you raise the issue at renewal.

Free: CRM health-check template

30 checks across integrations, automation, data quality and permissions. A good way to find the weak spots an outage exposes before the next one.

Get the template

How to check if Salesforce is down right now

When users say “Salesforce is down”, it is usually one of three things: a real Salesforce incident, a problem in your own org (a broken flow, a permission change, an expired integration credential), or a browser or network issue on the user’s side. Here is how to tell them apart quickly.

  1. Check Trust first. Open status.salesforce.com, search for your My Domain or instance name, and look for open incidents or maintenance.
  2. Subscribe to alerts. Trust lets you subscribe to notifications for your instance. Do it now, for your admin and one backup person, so you are not relying on users to tell you.
  3. Test from a second network and browser. If Salesforce works on mobile data but not in the office, the problem is probably local.
  4. Check scheduled maintenance. Release weekends (like the Winter ’27 upgrades) and planned maintenance can briefly limit access. Trust lists them in advance.
  5. Check your own recent changes. If only some users or one process is affected and Trust is green, look at recent deployments and Setup Audit Trail before blaming Salesforce.

Does Salesforce owe you credits for the outage?

We have not seen any report of Salesforce offering credits for this outage. Whether you are entitled to anything depends on what your contract says, so check your order form, your Master Subscription Agreement and your Success plan, and ask your account executive directly.

Even if there is nothing in the contract, an outage of this size is a fair point to raise at renewal, especially if you have your incident log and the business impact written down.

How to make your org less exposed next time

No customer can stop Salesforce from having an outage. What you can control is how much damage one does. These are the changes we would prioritise:

  • Make integrations safe to retry. Use upserts with external ID fields instead of plain inserts, so a retried call updates the same record instead of creating a duplicate.
  • Queue instead of failing. Put a message queue or middleware between Salesforce and critical systems, so data waits during an outage and flows in afterwards.
  • Give integrations their own users. Integrations running on a named employee’s login are hard to monitor and break when that person leaves. Salesforce is also moving AI agents to their own registered identities from November; our guide to what AIforce changes in your implementation covers that.
  • Monitor from outside Salesforce. A simple check that alerts you when logins or key API calls fail will tell you about the next outage before your users do.
  • Agree a fallback for the front line. Decide in advance how sales and support will work if Salesforce is unavailable for half a day: which shared inbox or spreadsheet they use, and who enters the data afterwards.
  • Write the runbook now. Turn the checklist above into a one-page document with names next to each step.

If you would rather have someone else watch your org, our Salesforce managed services team handles monitoring and post-incident clean-up, and our guide to Salesforce integration with other tools covers building integrations that survive this kind of failure.

Our take: the outage itself was fixed within a working morning in the US. The real cost for most companies will come from the quiet failures afterwards, such as a sync that skipped a night or a job that ran halfway. If you have not checked those yet, an hour spent on the checklist above is worth it.

Frequently asked questions

When was the Salesforce outage in September 2026?

It started on 16 September 2026 at 07:50 UTC (12:50 AM Pacific), during Dreamforce. Service was healthy from 15:26 UTC, and Salesforce marked the incident resolved at 18:59 UTC the same day.

What caused the Salesforce outage?

Salesforce said requests stalled while waiting on an internal login service, which used up server resources, and that a core system component was under increased load. It has not published a full root cause analysis as of 14 October 2026.

How many Salesforce instances were affected?

Salesforce described hundreds of instances in all regions. Reports tracking the Trust record put the final figure at 912 of about 3,243 instances. GovCloud was not affected.

Is Salesforce down today?

Check status.salesforce.com and search for your instance or My Domain. If Trust shows no incident, the problem is more likely in your org, your network or your browser.

Did Salesforce give credits for the outage?

No credits have been reported. Any entitlement depends on your contract and Success plan, so ask your account executive.

What should admins check after a Salesforce outage?

Scheduled jobs and scheduled flows, failed batch and queueable Apex, paused and failed flow interviews, integration errors, duplicates and gaps, web-to-lead and email-to-case volumes, and any support cases you tried to open during the outage.

How we put this guide together

The timeline and cause come from Salesforce’s Trust status updates for incident 20004433, as reported by Salesforce Ben, CIO and CX Today, cross-checked against each other. Where sources disagreed (for example on duration), we used the times Salesforce itself confirmed. The checklist reflects how we would approach recovery in a Salesforce org; the queries use standard Salesforce objects. We will update this page if Salesforce publishes a root cause analysis.

About the author

Ankur Pandey is Marketing Manager at Ashapura Softech, a certified Salesforce and Odoo partner with offices in Irving, Texas and Ahmedabad, India. He writes about Salesforce, Zoho and ERP planning for US mid-market businesses.

Categories
Agentforce Salesforce

Salesforce AIforce (Now SIForce): How It Impacts Your Salesforce Implementation

Last updated: 14 October 2026

Short answer

Salesforce AIforce does not move your data or replace your org. It is a layer that lets people and AI agents use Salesforce from Claude, Slack and Lightning, under the permissions you already have. What does change is how agents connect: each agent that calls Salesforce over MCP or API will need its own registered identity, and those calls get a new usage meter paid in Flex Credits. Salesforce is targeting November 2026 for these changes, and existing customers with API-based agents move over at renewal.

Three names in two months. Headless 360 in spring, AIforce at Dreamforce, and SIForce last week. If you look after a Salesforce org, the obvious worry is that one of these quietly turns into a migration project nobody budgeted for.

We read Salesforce’s own impact notes so you don’t have to, and the honest answer is: it’s smaller than it sounds, but it isn’t nothing. This guide walks through the AIforce impact on your Salesforce implementation, what stays exactly as it is, the dates that matter, and the checklist we’d work through before the new rules switch on. If you want the background on the announcement itself, start with our explainer What Is AIforce?

What is Salesforce AIforce?

AIforce is Salesforce’s interface layer for AI. It brings Salesforce data, workflows and permissions into the tools people already work in, so a rep can ask Claude or Slack about an account instead of opening Salesforce. Salesforce first showed the idea at TDX 2026 under the name Headless 360 and launched it as AIforce at Dreamforce on 15 September 2026.

It arrives in three forms, all running on the same foundation:

  • Claudeforce (Salesforce in Claude): a prebuilt MCP server with 37 sales skills that lets Claude read and act on Salesforce records. It launched in beta for Salesforce customers. Our guide to Claudeforce covers it in detail.
  • Slackforce: Salesforce context and actions inside Slack, such as creating accounts or logging notes from a prompt.
  • Agentforce Coworker: an AI teammate inside the Lightning interface that works within existing permissions and can call the Agentforce agents you have built. See our Agentforce Coworker guide.

Under all three sits the Headless Toolkit: the MCP servers, APIs, skills and plug-ins that expose the platform to outside AI tools. That toolkit is where the implementation impact comes from.

AIforce is now SIForce: what the rename means

On 9 October 2026, Marc Benioff posted on X: “AIForce is officially SIForce.” The rename followed a US executive order that asks federal agencies to say “super intelligence” instead of “artificial intelligence”.

For your org, the rename changes nothing yet. No features, prices or availability were announced with it, and Salesforce’s help documentation still uses the name AIforce. Our advice is the same as for the Agentforce product names Salesforce reversed in September: keep contracts, order forms and internal documents on the current names until your account team confirms how they will appear at renewal. In this guide we use “AIforce”, because that is the name in Salesforce’s own documentation today.

Do you need to migrate to AIforce?

No, not in the usual sense. There is no data migration, no new object model and no rebuild of your flows, reports or page layouts. AIforce reads from the org you already have. We already give every integration a dedicated, least-privilege user on the Salesforce CRM integration work we do, and agent identities follow the same idea.

The “migration” is narrower and more technical. It applies to agents: any AI tool, script or MCP client that connects to Salesforce on its own. Salesforce’s help article Understand AIforce Impact sets out two changes for them.

What Salesforce AIforce changes in your implementation

1. Salesforce agent registration: every agent gets its own identity

Right now, a lot of AI tools get into Salesforce by borrowing someone’s login or a shared integration user. Under the new model, admins register each agent through a feature called Agentic Identity. The agent gets its own identity on the platform, and you can give it a narrower permission set than any human user has.

Security-wise, this is a good change. An agent with its own identity can be limited, watched and switched off without touching a person’s account. The catch is the work behind it: someone has to find every agent, register it, and swap its OAuth credentials so Salesforce knows it’s talking to an agent and not a human.

2. Headless Platform Interactions: a new meter for agent calls

Each successful call a registered agent makes to Salesforce, over MCP or direct API, is metered as a Headless Platform Interaction (HPI). HPIs are tracked in Digital Wallet and paid for with Flex Credits, under the Customer 360 Platform section of the Flex Credits rate card.

Two details matter for budgeting:

  • The price isn’t set yet. The HPI multiplier is listed as “TBA”, and Salesforce says it will give 30 days’ notice before metering begins.
  • Your org must be on Flex Credit billing. Salesforce names Salesforce Foundations (listed at $0) or another SKU that enables Flex Credits. If Flex Credits are new to you, our guide to Agentforce pricing and Flex Credits explains how they work.

3. The deadline depends on how you buy and connect

Your situationWhat you must doDeadline
New purchase on or after 17 Sep 2026Register all agents that use Salesforce APIsWithin 3 months of Salesforce’s notice that Agentic Identity is available
Existing customer with API-based agentsRegister all agents and move them to the new billing modelAt your renewal
Any customer using a Salesforce MCP serverRegister the agents that use itWithin 3 months of the Agentic Identity notice
Traditional (non-agent) integrationsNothing: pricing and security options are unchangedNone

Salesforce is targeting November 2026 for agent registration, the new security controls and the billing model.

Our take: the rename gets the headlines, but the meter is what will show up in budgets. If your renewal falls in the next six months, it’s worth getting the agent inventory done now, while the multiplier is still “TBA” and nothing is being charged.

4. Sandboxes stay free

HPIs are only charged in active production orgs. Sandboxes, scratch orgs and Developer Edition orgs aren’t metered, so your team can build and test agents without using credits.

What stays the same

Most of your implementation doesn’t change at all. This is the part teams worry about most, so here it is side by side.

AreaChanges with AIforce?
Your data, objects and fieldsNo
Flows, Apex, validation rulesNo
Reports and dashboardsNo
Lightning UI and page layoutsNo: Lightning stays, and Agentforce Coworker runs inside it
Sharing rules and field-level securityNo, but they now control what AI can see (see below)
Traditional integrations (ERP sync, middleware)No: pricing and security options are unchanged
AI agents and MCP clients that call SalesforceYes: registration, new credentials and HPI billing
Contracts and product namesPossibly, at renewal: confirm with your account team

Where your current setup will show its gaps

AIforce doesn’t create new problems in an org. It makes the existing ones visible faster, because an AI answer is only as good as the data and permissions behind it.

  • Over-wide permissions. Every AIforce request runs as the user who makes it. If a profile has collected extra access over years of quick fixes, Claude or Slack will show that data to anyone who asks the right question. Permission sets that were “fine because nobody looks there” stop being fine.
  • Duplicate and stale data. An agent asked to brief a rep on an account reads whatever is in the org. Duplicate accounts and opportunities stuck in old stages produce answers that sound confident and are wrong. Approval steps catch bad writes, not bad reads.
  • Agents nobody listed. Scripts, browser extensions and MCP clients set up by individual teams are easy to miss. Under the new rules, an unregistered agent is both a security gap and an unplanned bill once metering starts.
  • Integrations running on a person’s login. If an integration or AI tool uses a named employee’s account, it breaks when that person leaves, and it hides what the tool is actually doing. Registration is the right moment to fix this.

If you want to measure these gaps before the November changes, our free CRM health-check template covers permissions, data quality and integrations in 30 checks.

AIforce readiness checklist for existing Salesforce orgs

Here’s the order we’d tackle it in. For most mid-sized orgs this is a few days of admin time, not a project. Start with the inventory; everything else depends on knowing what’s actually connected.

  1. Inventory every agent. List each AI tool, script and MCP client that connects to Salesforce, who owns it, and which user it runs as today.
  2. Split production from sandbox. Only production calls will be metered, so note which connections touch production.
  3. Design one permission set per agent. Give each agent the least access it needs. Don’t reuse an admin or a sales manager’s profile.
  4. Confirm Flex Credit billing. Check whether your org already has Salesforce Foundations or another SKU that enables Flex Credits.
  5. Note your renewal date. Existing API-based agents move to the new billing model at renewal, so that date sets your timeline.
  6. Plan the credential switch. Each agent will need new OAuth credentials tied to its registered identity. Schedule the reconnection so nothing breaks mid-quarter.
  7. Clean the data your first use case touches. Merge duplicate accounts and close out stale opportunities on the objects your pilot will read.
  8. Pilot read-only first. Start one team on Salesforce in Claude or Agentforce Coworker with read access, check the answers, then allow writes with approval.
  9. Watch for Salesforce’s notices. The Agentic Identity announcement starts the three-month clock, and the 30-day HPI notice tells you when billing begins.

Free: Salesforce implementation checklist

A step-by-step list covering permissions, data and integrations. Use it alongside this AIforce checklist before the November changes.

Get the checklist

Frequently asked questions

AIforce vs Agentforce: is it the same thing?

No. Agentforce is the set of AI agents that do work inside Salesforce. AIforce is the layer that lets people and agents reach Salesforce from Claude, Slack and Lightning. Agentforce Coworker is where the two meet. Our comparison of Claudeforce vs Agentforce goes deeper.

Is SIForce a different product from AIforce?

No. SIForce is the new name Marc Benioff announced on 9 October 2026 for the same product. No feature or pricing change came with it.

Do we have to pay extra for AIforce?

Salesforce hasn’t published a single AIforce price. What is confirmed is that registered agent calls to production orgs will be metered as Headless Platform Interactions and paid in Flex Credits, at a rate still listed as “TBA”. Confirm your entitlements with your account executive.

Will our existing integrations be billed as HPIs?

Traditional integrations keep their current pricing. The new meter applies to AI agents and MCP clients that are registered under Agentic Identity.

Which Claude plan does Salesforce in Claude need?

Published sources disagree, so check the current admin setup guide on AgentExchange or ask your Salesforce account team before you budget for it.

When do the new rules start?

Salesforce is targeting November 2026. New purchases from 17 September 2026 and anyone using a Salesforce MCP server have three months from Salesforce’s Agentic Identity notice to register agents. Existing customers with API-based agents move over at renewal.

The bottom line

AIforce, now SIForce, is not a migration of your Salesforce org. It is a new front door, and the house behind it stays the same. The work is in the doorway: knowing which agents come in, giving each one its own key, and budgeting for the new meter before it starts.

Ashapura Softech can run an AIforce readiness review for your org, covering an agent inventory, a permission audit and a Flex Credit estimate. Our Salesforce Agentforce consulting team handles agent setup and permissions, or you can talk to our Salesforce team directly.

How we put this guide together

Everything about registration, billing and deadlines here comes from Salesforce’s own help article on AIforce impact, dated 17 September 2026. Product details come from Salesforce’s Dreamforce announcements, and the rename from Marc Benioff’s post on 9 October. A few things are still open (the HPI price, final packaging), and we’ve said so where they come up rather than guessing. We’ll update this page when Agentic Identity goes live.

About the author

Ankur Pandey is Marketing Manager at Ashapura Softech, a certified Salesforce and Odoo partner with offices in Irving, Texas and Ahmedabad, India. He writes about Salesforce, Zoho and ERP planning for US mid-market businesses.

Categories
Agentforce Salesforce

Einstein Bots to Agentforce: Complete 6-Step Migration Guide

Einstein Bots to Agentforce is now a documented upgrade path: if your org still runs an Einstein Bot, Salesforce offers a faster way to start the move. At Dreamforce 2026, the company listed an Einstein Bots upgrade among its announcements, described as generally available. The idea: convert an existing bot into an Agentforce agent instead of rebuilding it from a blank screen.

This post covers what Salesforce documents about the conversion, what you should still plan to do by hand, and a cautious way to approach it. Where Salesforce has not published a detail, we say so. Use it as a planning guide for any Einstein Bots to Agentforce project.

In this guide

Einstein Bots to Agentforce: what Salesforce actually announced

The announcement page says you can click Upgrade Agent on Agentforce Labs to convert an existing Einstein Bot into an Agentforce Service Agent or Help Agent. The Help documentation adds that the current bot keeps running while you review and adjust the new agent.

That last point matters. This is not a switch that flips your live bot off. The documented behavior is that the tool creates a separate agent, so the original bot stays available.

One status detail to watch

The Dreamforce page lists the feature as GA. However, the Salesforce Help article we opened, Create AI Agents From Einstein Bots (Beta), still labels the Upgrade Agent tool as a beta service governed by Beta Services Terms. Documentation often lags announcements, but we could not confirm which label is current for your org. Check the Help article and your release notes before you plan around either one.

Why this is happening: Einstein Bots is in Maintenance Mode

Salesforce Help states: “Starting August 17, 2026, Einstein Bots is in Maintenance Mode. Existing Einstein Bots continue to be supported and can be used after August 17, 2026. However, no new features or enhancements will be added, and we recommend using Agentforce for new AI agent implementations.” You can read it in the Create an Einstein Bot article.

In plain terms: your existing bots are supported and keep working, but they will not get new capabilities. We did not find an official end-of-life date in the pages we opened, so do not assume one exists or does not exist. Salesforce’s direction is clearly toward Agentforce, and that is the real reason to start planning.

What converts when you upgrade a bot to Agentforce

According to the Help documentation, in an Einstein Bots to Agentforce conversion the Upgrade Agent tool analyzes your bot configuration and generates matching Agentforce Service Agent components:

  • Dialogs become subagents, based on grouped dialogs with similar intents.
  • Action steps become agent actions.
  • Intents are carried into the new agent structure.

A second Help article, Upgrade Einstein Bot to Agentforce Service Agent, describes the input more broadly as bot dialogs, intents, dialog steps, variables, entities and training data. Treat that as the scope the tool reads from, not a promise that every element comes across intact.

Einstein Bot elementBecomes in the Agentforce Service Agent
Dialogs (grouped by similar intent)Subagents
Action stepsAgent actions
IntentsCarried into the new agent structure
Flows, Apex, integrationsReuse is possible, but must be reviewed and tested

The first three rows of this Einstein Bots to Agentforce mapping come from Salesforce Help. The last row is our planning advice, not a Salesforce statement.

What does not convert, or is unclear

This is where you need to be careful. The Help article we reviewed does not list what fails to convert or any limitations, so treat the list of exclusions as unknown until you test a conversion yourself. Salesforce tells you to review and configure the new agent while the old bot stays live, so plan to test thoroughly before you switch any traffic.

Some community writers are more specific. The Concret migration guide argues that dialog trees, NLU training data and named conversation variables have no direct Agentforce equivalent, while Flows, Apex classes and external integrations can usually be reused with minor updates. A second community article makes a similar case. These are author opinions, not Salesforce statements, so use them as a prompt for your own review rather than as fact.

The practical lesson for Einstein Bots to Agentforce projects is the same either way: expect the conversion to give you a draft, not a finished agent.

Prerequisites for converting Einstein Bots to Agentforce

For Einstein Bots to Agentforce conversions, the Salesforce Help pages above list:

  • Editions: Enterprise, Performance, Unlimited, or Developer Edition with Foundations or Agentforce 1 Editions. Enhanced Messaging channels have their own edition requirements.
  • Interface: Lightning Experience.
  • An existing Einstein Bot in a connected org, production or sandbox.
  • Bot permissions: Customize Application, or Modify Metadata, or Manage Bots.
  • Agent permissions: Manage Agentforce Service Agents and Manage AI Agents, plus Agentforce Default Admin or Customize Application.

The Help article describes two ways to run the conversion: the Agentforce Labs web tool, or a Claude Code route that uses Salesforce CLI (version 2.0.0 or later) and Salesforce’s published skills. With the web tool, a production org is added automatically, while a sandbox must be connected manually. You can also upload a bot package.

The pages we opened did not mention Data 360 (formerly Data Cloud) requirements for the conversion itself, so confirm this with your Salesforce account team rather than assuming either way.

What the Agentforce agent may cost

Salesforce does not appear to publish a price for the conversion tool itself. The cost question is about running the agent afterward. On its Agentforce pricing page, Salesforce lists Flex Credits at $500 per 100,000 credits, and also lists a per-conversation option at $2 per conversation. Prices and packaging change, so confirm current terms for your contract.

For any Einstein Bots to Agentforce plan, the key point for ops leaders is that usage-based models behave differently from a bot you have run for years. Model your real monthly conversation volume before you commit. If you are weighing options more broadly, our comparison of Claudeforce vs Agentforce may help frame the trade-offs.

Einstein Bots to Agentforce: a safe migration approach

Infographic: 6-step Einstein Bots to Agentforce migration path
Six steps to migrate Einstein Bots to Agentforce safely.

1. Inventory your Einstein Bots before you convert

List every active bot, its dialogs, the Flows and Apex it calls, the channels it serves and its monthly volume. Find the dialogs that carry most of the traffic.

2. Convert in a sandbox first

Run the tool against a sandbox copy, and treat the first Einstein Bots to Agentforce run as a rehearsal. Because the original bot stays live, you can inspect the generated agent in Agentforce Builder without touching customers.

3. Review the Einstein Bots to Agentforce draft line by line

Check the generated subagents, actions and instructions against what your bot really does. Look for anything that did not carry across, and rebuild it deliberately.

4. Test with real conversation examples

Use anonymized transcripts from your own history. Test edge cases, escalation to a human, and anything involving sensitive data.

5. Consider a hybrid handoff

You do not have to cut over all at once. Salesforce documents how to transfer conversations from an enhanced bot to an Agentforce Service Agent using an Omni-Channel flow with a Route Work action. A Salesforce developer blog describes using the bot as a front-line router that passes complex inquiries to specialized agents. That lets you move one topic at a time.

6. Activate gradually

Turn the new agent on for one channel or one topic first, watch the results, and keep the old bot available until you are confident.

Release timing also matters. If you are lining up changes with your release calendar, see our notes on the Salesforce Winter ’27 release.

Einstein Bots to Agentforce pre-migration checklist

  • Confirm your edition includes Foundations or Agentforce 1.
  • Confirm the required bot and agent permissions are assigned.
  • Document each bot’s dialogs, Flows, Apex and channels.
  • Pull monthly conversation volume and escalation rates.
  • Check the current status (beta or GA) of the Upgrade Agent tool.
  • Ask Salesforce about Data Cloud and pricing for your contract.
  • Pick one low-risk, high-volume topic to start with.
  • Run the first conversion in a sandbox.
  • Define human escalation and test it.
  • Decide who signs off before activation.

Einstein Bots to Agentforce: frequently asked questions

Does the conversion replace my live bot?

No. Salesforce says the tool creates the new agent separately, so your existing bot remains available while you review and configure the new agent. Plan to test the new agent before you route any live traffic to it.

Is Einstein Bots being shut down?

Not according to the pages we opened. Salesforce says Einstein Bots entered Maintenance Mode on August 17, 2026: existing bots remain supported, but no new features or enhancements are coming. We found no official end-of-life date.

What do I need before I can convert a bot?

You need an Enterprise, Performance, Unlimited or Developer Edition org with Foundations or Agentforce 1, Lightning Experience, an existing bot, and the bot and agent permissions listed above.

Will everything in my bot carry over?

Do not assume so. Dialogs, action steps and intents are converted, but Salesforce’s Help article does not list what is left out, so we could not confirm exclusions. Plan to review and rebuild some pieces manually.

How much will the new agent cost?

The conversion tool’s cost is not stated in the sources we opened. Running the agent is priced separately; Salesforce lists Flex Credits at $500 per 100,000 credits and a $2 per conversation option. Confirm current pricing for your contract.

Getting help with the move

Einstein Bots to Agentforce conversion is the easy part. Deciding what the agent should do, how it hands off to people, and what it will cost to run takes planning. If you want a second set of eyes, Ashapura Softech can help through our Salesforce Agentforce consulting services, and you can contact us to talk it through. If you are still choosing a partner, our Agentforce implementation partner guide explains what to look for.

Categories
Salesforce Agentforce

Salesforce Winter ’27 Release: 15 Critical Changes to Test

Last updated: 15 October 2026

Short answer

Salesforce Winter ’27 is now live in every production org. Salesforce upgraded production instances in waves on 4 September, 2 October and 9 October 2026, and the release reached general availability on 12 October. If you have not tested yet, start with the changes that can stop integrations: the new API permission for SOAP login(), the retirement of the OAuth username-password flow on 20 February 2027, and the end of standard-volume platform events in Summer ’27.

The Salesforce Winter ’27 release went generally available on October 12, 2026, and it is not a quiet one. Alongside the usual Flow and admin improvements, it carries security changes that can stop integrations from logging in, a platform events retirement with a real deadline, and a large set of Agentforce updates that change how agents are built and measured.

This guide lists the 15 Salesforce Winter ’27 release changes we think Salesforce admins, architects and IT leads should actually test, in order of risk. It is written for teams running a live production org who want to know what to check now that their org is on Winter ’27, not for people who want to read every line of the release notes. For each item you will find what changed, who it affects and the specific test to run.

We work on Salesforce orgs every week as an implementation partner, so the list leans towards what tends to break in real orgs rather than what looks best in a keynote.

Related: Salesforce also had a major outage on 16 September 2026. See what happened in the Salesforce outage and what admins should check.

Salesforce Winter ’27 release dates at a glance

Running an Einstein Bot? Read our guide to converting Einstein Bots to Agentforce agents, which Salesforce announced at Dreamforce 2026.

Planning to adopt new releases? Our guide to CRM implementation services covers what a rollout usually costs and how long it takes.

Salesforce rolls a major release out in waves, so your production org may upgrade on a different weekend from someone else’s. The dates below are the ones that matter for planning. Always confirm your own maintenance window on the Salesforce Trust status site for your instance.

Salesforce Winter '27 release dates timeline with GA on October 12, 2026 and 2027 retirement deadlines
MilestoneDateWhat it means for you
Pre-release orgs openAugust 13, 2026Free trial orgs with Winter ’27 features, but none of your own configuration
Release notes publishedAugust 19, 2026Full list of features and release updates available to read
Sandbox previewLate August 2026Preview sandboxes upgraded early, the best place to test your own setup
Production upgrades4 September, 2 October and 9 October 2026Your date depends on your instance, check Trust status
General availabilityOctober 12, 2026All production orgs on Winter ’27
OAuth username-password flow retiredFebruary 20, 2027Integrations using grant_type=password stop getting tokens
SOAP login() call retiredJune 1, 2027Integrations that log in through SOAP must move to OAuth
Standard-volume platform events retiredSummer ’27Events must be migrated to high-volume before then

Free CRM health check: 30 checks in 5 pages

Score your CRM in about an hour across data quality, pipeline, adoption, automation, reporting and security. Free PDF, no signup.

Download the free CRM health check

Planning Agentforce? Start with our free Agentforce readiness checklist, or book a free 30-minute consultation.

Already on Winter ’27? Your first-week checks

If your org upgraded in one of the October waves and nothing has broken yet, that is a good sign, but several of the riskiest changes only bite later or only on specific integrations. These are the checks we run first for clients after a major release lands:

  1. Integration users that call SOAP login(). Confirm each one has the new API permission described in change 1 below. A failed nightly sync is often the first sign this was missed.
  2. Anything using the OAuth username-password flow. List every connected app and middleware job that still sends a username and password for a token. They keep working until 20 February 2027, then stop. Moving them to a supported OAuth flow is a project, not a setting, so plan it now.
  3. Standard-volume platform events. Pull a list of the events your org publishes and subscribes to, and mark which are standard-volume. They have to move to high-volume before Summer ’27.
  4. Your three most important flows. Re-run them in the new Flow Test Mode (change 6) with real-looking records, so you know they behave the same after the upgrade.
  5. Agents in production. Check that existing Agentforce agents still pick the right actions after the skills and plugins registry changes (change 11).

Looking further ahead, Salesforce has said its CRM reasoning model, Koa, should reach general availability in winter 2026 in US regions. Our guide to Salesforce Koa explains what it is and how to get your agents ready.

Salesforce Winter ’27 known issues: what admins are running into

Most of the problems we have seen since the October upgrade weekends are not bugs. They come from features that changed during the release, beta features that were switched off, and behaviour that only shows up with real users. Before you raise a case, check these first.

  • Check the official Known Issues site. Search issues.salesforce.com for the feature name and subscribe to any issue that matches. Salesforce posts workarounds there before it posts fixes.
  • Features that changed mid-release. Flow tags were removed the week of 14 September (see below). Reactive formulas in conditional visibility rules were pulled on 14 September and then added back to the notes in early October. If a feature you read about in a preview blog is missing, check the release notes change log before you assume your org is broken.
  • Flow record-lock retry is narrower than it sounds. The automatic retry on locked records applies only to schedule-triggered flows on run-time API version 62.0 or lower, with one retry after 10 seconds. It is not a general fix for lock errors in record-triggered flows (release note).
  • Retired betas. Teams that relied on beta features have reported that some, such as Work Summaries for Case and the Experience Builder agent, no longer work. Move users to the generally available alternative rather than waiting for the beta to return (Maintask summary).
  • Agentforce appearing in Setup. Partners report that eligible orgs are having Agentforce enabled at no extra cost. Enabled is not the same as deployed: no agent runs until someone builds and activates it, so review who has the Manage AI Agents permission (CETDIGIT).
  • “It works for me” reports. Profile Filtering and the 200% zoom changes only show up for non-admin users. Always reproduce a reported issue while logged in as the affected user.

How this Salesforce Winter ’27 release list is organised

The 15 Salesforce Winter ’27 release changes fall into three groups: five that can break things, five Flow and admin wins and five Agentforce updates. The first five can break something that works today, so they come first and deserve testing time before your upgrade weekend. The next five are Flow and admin improvements that are worth adopting because they save time or reduce risk. The last five are Agentforce and cloud updates that mostly affect planning and budgets rather than day to day stability.

If you only have one afternoon, test group one. If you have a week, work through all three.

Part 1: Five Salesforce Winter ’27 release changes that can break things

1. SOAP login() needs a new API permission

The Salesforce Winter ’27 release enforces a release update titled “Assign Use Any API Auth Permission for SOAP login()”. Users who authenticate through the SOAP API login() call now need that permission assigned, either through a profile or a permission set.

Who it affects: any integration user, middleware tool or older data loader setup that still logs in through SOAP rather than OAuth. These are often integrations nobody has looked at in years.

What to test: in your preview sandbox, list every integration user and check which ones authenticate through SOAP login(). Assign the permission through a permission set rather than editing profiles, then run each integration once end to end. Also note the bigger deadline behind this change: the SOAP login() call itself retires on June 1, 2027, so this is the moment to plan the move to OAuth rather than patching it twice.

2. Profile Filtering is switched on by default

With Profile Filtering enforced by the Salesforce Winter ’27 release, users without the View All Profiles permission can no longer see the names of profiles they are not assigned to. On its own this is a sensible security change. The problem is anything built on the assumption that profile names are visible.

Who it affects: flows, validation rules, formula fields, Apex and reports that read Profile.Name for users who do not hold View All Profiles, plus any custom user management screens.

What to test: search your metadata for references to Profile.Name, run the affected automation as a standard user in the sandbox and confirm it still behaves. Where logic depends on profile, consider switching it to a custom permission, which is the more durable pattern anyway.

3. The OAuth username-password flow is retired on February 20, 2027

This is the change most likely to cause a silent outage. Salesforce is retiring the OAuth 2.0 username-password flow for connected apps, where an integration sends a username, password and security token to get an access token. Enforcement is set for February 20, 2027. When it lands, those integrations simply stop getting tokens, and there is no warning screen for a user to notice.

Who it affects: existing orgs with server to server integrations still using grant_type=password. New orgs already have the flow blocked by default.

What to test: ask every integration owner which OAuth flow they use. Move nightly jobs and middleware to the client credentials flow running as a dedicated integration user, or to the JWT bearer flow for more sensitive connections. Store secrets in Named Credentials, not in code or config files. The Salesforce Winter ’27 release is the moment to do this, because February leaves little room once year end freezes are taken into account.

4. Standard-volume platform events retire in Summer ’27

You can no longer define new standard-volume custom platform events, and new platform events are high-volume by default. Existing standard-volume events are due to retire in Summer ’27, according to the Salesforce Platform Events Developer Guide.

Who it affects: orgs with older event-driven integrations, especially ones built several years ago when standard volume was the default.

What to test: list your custom platform events and check each one’s publish behaviour. For every standard-volume event, plan the migration to high-volume and retest the subscribers, including any Apex triggers, flows and external CometD clients. The migration is not hard, but it touches every subscriber, so it belongs in a planned release rather than a rushed fix.

5. Accessibility updates change layouts above 200% zoom

Three accessibility updates enforced by the Salesforce Winter ’27 release improve how cards, docked containers, menu lists, panels, date pickers, popovers, utility bars, record headers, page headers and modal windows behave at zoom levels above 200%.

Who it affects: most orgs will simply get a better experience. The risk sits with custom Lightning components and CSS overrides that assume a fixed layout.

What to test: open your most used record pages, custom components and utility bar items at 200% and 400% browser zoom in the sandbox. Look for clipped buttons, overlapping text and modals that cannot be closed. Fix those before users who rely on zoom run into them.

Winter ’27 is live

Will Winter ’27 break anything in your org?

We check your org against the release changes most likely to cause problems, so you can fix them before your users notice.

  • Authentication and integration changes that can stop logins or syncs
  • Flows and automations that need re-testing
  • A short, prioritised fix list your admin can act on

Book a Winter ’27 impact check

Certified Salesforce partner · 30-minute call, no obligation

Part 2: Five Flow and admin changes in the Salesforce Winter ’27 release

If your org still runs most of its day-to-day automation by hand, our rundown of the Sales Cloud workflows worth automating first pairs well with this section.

6. Flow Test Mode (beta) replaces Debug Mode

In the Salesforce Winter ’27 release, Test Mode gives Flow Builder a dedicated place to test a flow, combining what debug and flow tests used to do separately. You can save test scenarios with mock inputs and assertions, rerun them after every change and track coverage.

Why it matters: most flow failures in production come from a change to one branch breaking another branch nobody retested. Saved, repeatable tests are the cheapest protection against that.

What to do: pick your three most important record-triggered flows and build saved tests for them in the sandbox. Because Test Mode is still beta, keep your existing testing process running alongside it for now.

7. Screen flows can run on many records at once

After the Salesforce Winter ’27 release, a screen flow can run as a mass quick action from a list view or a related list, receiving the selected records as a collection of IDs. Screen flow quick actions also get custom modal sizing.

Why it matters: bulk updates that used to need a data loader job, a custom Lightning component or a developer can now be built by an admin.

What to do: look for requests your team handles with spreadsheets and data loader, such as reassigning a batch of cases or updating stages on selected opportunities. Those are good first candidates.

8. Flow Builder catches more mistakes at save time

With the Salesforce Winter ’27 release, Flow Builder warns when a Create Records element is missing required field assignments and catches field length violations when you save, instead of failing at runtime. There is also a new run option that makes a flow respect the running user’s permissions, plus an edit history panel and better version comparison.

Why it matters: fewer broken flows reach production, and you can see who changed which element and restore an older version.

What to do: open your largest flows in the sandbox after the upgrade and fix any new warnings. Then review which flows run in system context by default and decide whether they should run in user context instead. For guidance on getting an org ready for automation at this level, our Agentforce readiness checklist covers the permission and data clean up that makes this safer.

9. Apex heap limits go up

The Salesforce Winter ’27 release raises the Apex heap size limit from 6MB to 10MB for synchronous transactions and from 12MB to 25MB for asynchronous transactions. Developers can also test External Services and HTTP callouts in Apex integration tests (developer preview), and a new Apex Symbol API (beta) exposes Apex type metadata.

Why it matters: heap limit errors are a common cause of failed batch jobs and integrations that process large payloads. Some of those will now simply work.

What to do: check your logs for recent heap limit exceptions. Rerun those jobs in the sandbox, but do not treat the higher limit as permission to skip optimisation. Code that barely fitted in 6MB will not stay comfortable for long.

10. Reports get record preview and a home on LWR sites

The Salesforce Winter ’27 release lets you preview records straight from Lightning reports without leaving the report (beta), and Lightning reports and dashboards can now be embedded in Lightning Web Runtime Experience Cloud sites (beta).

Why it matters: partners and customers on an LWR portal can finally see real reports and charts, and internal users lose fewer clicks moving between a report and the records behind it.

What to do: if you run a partner or customer portal, check which reports your external users currently receive by email or export. Those are candidates for embedding once the feature leaves beta.

Part 3: Five Agentforce updates in the Salesforce Winter ’27 release

11. Agent Skills and Plugins get a unified registry

Salesforce’s Winter ’27 announcement introduces a unified registry for Agent Skills and Plugins with more than 100 prebuilt skills. Developers can create reusable, governed capabilities that work across surfaces instead of rebuilding the same action for each agent.

Why it matters: the cost of an Agentforce rollout is often in building and maintaining actions. A shared, governed library changes that equation, and it follows the same direction as the seven Agentforce named agents announced in September.

What to plan: list the actions your agents use today and check which ones now exist as prebuilt skills. Retiring custom actions you no longer need is also a maintenance saving.

12. Custom Scorers (beta) for agent quality

The Salesforce Winter ’27 release adds Agentforce Custom Scorers in beta, letting you define your own evaluation logic for agent sessions on top of the built-in quality metrics.

Why it matters: a generic quality score cannot tell you whether an agent followed your refund policy or escalated at the right point. Custom scoring lets you measure the rules that matter to your business.

What to plan: write down three or four rules every agent conversation must follow, and turn those into scorers in a sandbox. This is also the evidence a steering committee will ask for before approving wider rollout.

13. Adaptive Experiences and Dynamic Plans in service

Adaptive Experiences and Dynamic Plans listen to service conversations in real time and generate and update a resolution plan as the issue evolves. Salesforce says four customers are live, including PowerSchool with more than 550 users.

Why it matters: this moves AI from summarising after the call to guiding the agent during it, which is where handle time actually drops.

What to plan: this depends heavily on clean case data and well structured knowledge. If your knowledge base is out of date, fix that first. For consumption-based costs, our explainer on Agentforce pricing and Flex Credits shows how usage is billed.

14. Agentforce Contact Center expands, with autonomous scheduling

In the Salesforce Winter ’27 release, Agentforce Contact Center expands to more than 30 countries, bringing voice, digital channels and CRM together. Autonomous scheduling with Agentforce Voice lets an agent check availability and book or dispatch technicians around the clock, turning a booking call that used to take around 15 minutes into seconds.

Why it matters: field service and appointment-heavy businesses get a practical use case with a clear return, rather than a general chatbot.

What to plan: map your current booking calls. If most follow the same pattern of checking availability and confirming a slot, they are good candidates. Check your edition too, because what is included changed with the new Salesforce Core, Advanced and Max editions.

15. Agentforce orchestrates third-party agents

With the Salesforce Winter ’27 release, Agentforce can orchestrate agents running on AWS, Azure and Google through A2A-compliant ecosystems. In other words, a Salesforce agent can hand work to an agent built elsewhere and receive the result back.

Why it matters: few large companies will run every agent on one platform. Orchestration across vendors is what keeps a multi-vendor setup from turning into disconnected bots.

What to plan: if you already run agents outside Salesforce, list them and the data each one needs. Governance, logging and permissions across vendors will need an owner before this goes live. If you are weighing Salesforce’s own agents against Claude inside Salesforce, our comparison of Claudeforce vs Agentforce covers the trade-offs.

Smaller Sales and Service changes in the Salesforce Winter ’27 release

A few smaller Salesforce Winter ’27 release updates will not break anything but are worth showing users. In Sales, Einstein Conversation Insights suggests follow-up actions after meetings, reps can capture voice notes after a meeting in the Salesforce mobile app, and Pipeline Forecasting shows deal risks, methodology scores and contact insights. In Service, agents can view original case attachments inline in case details, delete outdated milestones from the case timeline, and use a revamped Enhanced Case Merge interface (beta).

What was pulled from the Salesforce Winter ’27 release: Flow Tags

Flow Tags, which would have let admins label and group flows by category, was removed from the Winter ’27 release during the week of September 14, 2026. The release notes change log records the removal, and Salesforce said only that the feature is not quite ready yet.

The practical lesson is simple. Plan and budget on what is generally available, not on what appeared in a preview, a keynote or an early version of the release notes. Keep using naming conventions and descriptions to organise flows until Flow Tags returns.

Salesforce Winter ’27 release testing checklist

Every Salesforce Winter ’27 release test needs an owner: integration leads cover authentication, admins cover Flow and profiles, and developers cover Apex and platform events. The table maps each change to its test and usual owner.

#AreaWhat to testUsual owner
1SOAP login()Assign the new API auth permission, run each SOAP integration end to endIntegration lead
2Profile FilteringFind Profile.Name references, run them as a standard userAdmin
3OAuth flowsList integrations on grant_type=password, plan client credentials or JWTIntegration lead
4Platform eventsList standard-volume events, plan migration to high-volumeDeveloper
5AccessibilityCheck key pages and custom components at 200% and 400% zoomAdmin, UX
6Flow Test ModeBuild saved tests for the three most important flowsAdmin
7Bulk screen flowsReplace one data loader task with a mass quick actionAdmin
8Flow save checksOpen large flows, clear new warnings, review run contextAdmin
9Apex heapRerun jobs that hit heap limits recentlyDeveloper
10ReportsIdentify portal reports suited to LWR embeddingAdmin
11 to 15AgentforceMap actions to prebuilt skills, draft custom scorers, list external agentsProduct owner

Testing Winter ’27? Check the rest of your org too

Release week is a good time to find what else is broken. Our free 30-point CRM health check covers data, automation, reporting and access.

Download the free CRM health check

How to run the Salesforce Winter ’27 release upgrade in your org

  1. Confirm your date. Look up your production instance on Salesforce Trust status and put the maintenance window in the team calendar.
  2. Test in a preview sandbox, not a pre-release org. Pre-release orgs have the new features but none of your configuration, so they cannot tell you what breaks.
  3. Start with the release updates. Open Setup, then Release Updates, and work through anything marked for enforcement before touching new features.
  4. Run your regression tests. Apex tests, key flows, integrations and the top ten record pages your users open every day.
  5. Tell users what changes for them. A short note on the few visible changes saves a week of support tickets.
  6. Book the follow-up work. The OAuth, SOAP and platform event deadlines fall in 2027, but the migrations need a slot in your plan now.

Teams without the capacity to test every release properly often hand this recurring work to Salesforce managed services.

Frequently asked questions about the Salesforce Winter ’27 release

When is the Salesforce Winter ’27 release date?

Salesforce announced general availability of the Salesforce Winter ’27 release on October 12, 2026. Production orgs upgrade in waves before that date, and the exact weekend depends on your instance, which you can check on Salesforce Trust status.

What breaks in Salesforce Winter ’27?

The Salesforce Winter ’27 release changes most likely to cause problems are the SOAP login() permission requirement, Profile Filtering being on by default and accessibility changes affecting custom components. Two later deadlines matter just as much: the OAuth username-password flow retires on February 20, 2027 and standard-volume platform events retire in Summer ’27.

Is Flow Test Mode generally available in Winter ’27?

No. Flow Test Mode is in beta in the Salesforce Winter ’27 release. It is safe to use in a sandbox for building saved flow tests, but keep your existing testing process alongside it until it reaches general availability.

Did Salesforce remove Flow Tags from Winter ’27?

Yes. Flow Tags was removed from the Salesforce Winter ’27 release in September 2026, with the change recorded in the release notes change log. Salesforce has not given a new date.

What should Salesforce admins test first in Winter ’27?

Test the enforced release updates first: SOAP login() authentication, Profile Filtering and the accessibility changes. Then check which integrations still use the OAuth username-password flow, because those will stop working on February 20, 2027.

Do Winter ’27 Agentforce features cost extra?

It depends on the feature and your edition. Check the availability section of each feature in the Salesforce Winter ’27 release notes before planning a rollout. Many Agentforce capabilities consume Flex Credits, and edition packaging changed in September 2026 with Core, Advanced and Max.

Getting help with the Salesforce Winter ’27 release

Ashapura Softech is a Salesforce implementation partner working with US businesses on Salesforce CRM development, Salesforce cloud services and CRM software development, and builds the custom middleware behind many Salesforce integrations through our enterprise software development team. If you want a second pair of eyes on your release updates, your integration authentication or an Agentforce rollout, we are happy to help, and you can see how we approach delivery in our Agentforce implementation partner guide.

Winter ’27 is live

Will Winter ’27 break anything in your org?

We check your org against the release changes most likely to cause problems, so you can fix them before your users notice.

  • Authentication and integration changes that can stop logins or syncs
  • Flows and automations that need re-testing
  • A short, prioritised fix list your admin can act on

Book a Winter ’27 impact check

Certified Salesforce partner · 30-minute call, no obligation

Categories
Agentforce Salesforce

Agentforce Coworker: What It Does, What It Costs and How to Turn It On

Of everything Salesforce announced under AIforce at Dreamforce 2026, Agentforce Coworker is the piece most admins will meet first. It is not a pilot you apply for. In many orgs it is already switched on, because Salesforce began turning it on automatically from 10 August 2026. This guide explains what Agentforce Coworker is, what it can and cannot touch, what it costs, how to turn it on or keep it off, and how to fix the errors admins run into during setup.

Related: Agentforce Coworker is one of the three AIforce interfaces. If you are planning for the new agent registration and billing rules, read how Salesforce AIforce impacts your implementation.

Agentforce Coworker at a glance

QuestionShort answer
What it isAn AI assistant in the Salesforce search bar that answers questions in plain language and can act on CRM records
Where it worksSalesforce, Slack, Microsoft Teams, Claude, ChatGPT and mobile
What it readsSalesforce CRM, Slack and connected Data 360 sources
EditionsEnterprise, Unlimited and Agentforce 1
CostNo extra fee for CRM and Slack search on eligible Unmetered seats; Data 360 answers use credits
StatusAnnounced May 2026, automatic enablement from 10 August 2026, part of AIforce since 15 September 2026

Free CRM health check: 30 checks in 5 pages

Score your CRM in about an hour across data quality, pipeline, adoption, automation, reporting and security. Free PDF, no signup.

Download the free CRM health check

Related: Coworker works with your existing Agentforce agents, and from winter 2026 those agents can run on Salesforce Koa, Salesforce’s new CRM reasoning model.

Planning Agentforce? Start with our free Agentforce readiness checklist, or book a free 30-minute consultation.

What is Agentforce Coworker?

Agentforce Coworker is an AI assistant that sits in the Salesforce search bar inside Lightning. A user types a question in plain language, and instead of a list of matching records they get one written answer built from the data they are allowed to see. Salesforce’s own Coworker FAQ describes it as a teammate that can reason over complex requests, build a multi step plan and call backend tools to finish the task.

The simplest way to think about it: Coworker replaces “search, open five records, read, piece it together” with “ask once, get the summary”. It does not replace your agents, your flows or your reports. It sits in front of them and becomes the single entry point to them.

Agentforce Coworker timeline

  • May 2026: Marc Benioff announced Coworker as an AI teammate “inside every Salesforce search bar”, and it opened as a beta for Agentforce customers.
  • 10 August 2026: Salesforce started enabling it automatically for eligible orgs, on a rolling basis.
  • 15 September 2026: At Dreamforce it became one of the three products inside AIforce, next to Claudeforce and Slackforce. Salesforce said 100,000 users activated it in the first 35 days.
Agentforce Coworker timeline: beta in May 2026, automatic enablement from 10 August 2026, part of AIforce from 15 September 2026
Agentforce Coworker timeline in 2026.

Agentforce Coworker vs traditional Salesforce search

Classic global search is good at one thing: finding a record when you already know roughly what it is called. Coworker is built for the questions that sit between records.

Traditional Salesforce searchAgentforce Coworker
InputKeywords and record namesPlain language questions
OutputA list of matching recordsOne written answer with the records behind it
SourcesSalesforce objectsSalesforce, Slack and Data 360 sources
ActionNone, the user opens and edits recordsCreates or updates CRM records after confirmation, routes work to agents
SecurityUser permissionsUser permissions plus Data 360 governance, masking and grounding
Agentforce Coworker vs traditional Salesforce search: input, output, sources, actions and security compared
Agentforce Coworker compared with traditional Salesforce search.

Where Agentforce Coworker works

The Lightning search bar is where most users first see it, but Salesforce’s Agentforce Coworker page lists more surfaces: Slack, Microsoft Teams, Claude, ChatGPT and mobile. The point is that the same assistant, with the same permissions, follows the user into whichever tool they already have open. That is the idea behind AIforce as a whole, and it is why Coworker pairs naturally with Claudeforce for teams that already work in Claude.

What Agentforce Coworker can search and do

According to the Salesforce developer guide and the help FAQ, Coworker reads across three kinds of source:

  • Salesforce CRM data: accounts, contacts, opportunities, cases, custom objects, knowledge articles and past case history.
  • Slack: messages, channels and conversations the user can already access.
  • Data 360 sources: Salesforce says more than 270 external sources connect through Data 360 connectors, including data lakes, knowledge bases and enterprise apps such as Jira.

On top of search, it does three things that matter in daily work:

  • Create or update records. It checks permissions and validation first, and asks the user to confirm before it saves anything. Today this works only on Salesforce CRM records, not on Slack or external data.
  • Route the request. If a question belongs to a specialised agent, for example a service agent or one of the named agents like Casey or Piper, Coworker hands it over rather than guessing.
  • Summarise across sources. A sales rep can ask what is putting a deal at risk and get the open cases, the Slack thread about pricing and the stalled approval in one answer.
How Agentforce Coworker answers a question: user question, permission check, CRM, Slack and Data 360 sources, answer or action
How Agentforce Coworker turns one question into an answer or an action, inside the user’s own permissions.

Agentforce Coworker use cases and example prompts

The value shows up when people ask the questions they currently answer by opening several tabs. These are the kinds of prompts that fit each team.

Sales

  • “What should I know before my call with this account tomorrow?”
  • “Which of my opportunities closing this quarter have had no activity in two weeks?”
  • “Log a note on this opportunity: pricing approved, legal review next week.”

Service

  • “Has this customer reported this issue before, and how was it fixed?”
  • “Why are cases about login errors spiking this week?”
  • “Find the knowledge article that answers this case.”

Customer success and leadership

  • “Summarise open cases, renewal date and recent Slack discussion for this account.”
  • “Which strategic accounts have an open escalation and a renewal in the next 90 days?”

For the wider picture of where agents earn their keep, see our Agentforce use cases guide.

How it handles security and permissions

Salesforce says Coworker “fully honors the existing Salesforce access model”. It runs with the permissions of the person using it, so if a user cannot open a record, Coworker cannot read it for them. Behind that sit attribute based access control, PII masking, toxicity detection and result grounding through Data 360, and the AIforce announcement repeats the zero data retention model for the underlying language models.

External sources need more thought. Permission aware sources carry their access rules across automatically, and federated sources enforce permissions in real time at the source. For other connectors, an admin has to create governance policies in Data 360 by hand. Data from sources that are not zero copy is ingested and stored in Data 360, which matters for data residency reviews.

The practical catch is that “respects your permissions” is only as good as the permissions themselves. Many orgs have years of broad sharing rules, public read on objects that should be private, and profiles that were cloned instead of designed. Search used to hide that problem because nobody scrolled through everything. A good AI answer surfaces it in seconds. If you have not reviewed sharing in a while, do that before you roll Coworker out widely. Our Agentforce readiness checklist covers the same ground.

Automatic enablement: is Coworker already on in your org?

Salesforce’s automatic enablement FAQ says the rollout began on 10 August 2026 on a rolling basis. Your org is in scope if it has Unmetered User Based AI seats and Einstein is active.

Some orgs are left out of automatic enablement:

  • Orgs that have opted out of generative AI.
  • Multi dataspace or multi org setups.
  • HIPAA, health and life sciences, and government orgs.
  • Flex Foundation seats without Unmetered entitlements.

If your org is in scope and you would rather roll it out on your own schedule, there are two ways to opt out: the link in the notification email Salesforce sends, or Setup > Agentforce Coworker > Opt Out of Auto Activation. Users who gain an eligible seat after the automatic switch on still need an admin to give them access by hand.

Coworker rollout review

Coworker may already be switched on. Is your data ready for it?

Coworker answers from whatever your users can already see. We review permissions and data quality first, so it gives the right answers to the right people.

  • Check whether Coworker is enabled in your org today
  • Find profiles and sharing rules that expose more than they should
  • A simple rollout plan for your first team

Get a Coworker rollout review

Certified Salesforce partner · 30-minute call, no obligation

Agentforce Coworker pricing: what does it cost?

For most seat based customers, the search itself is included. Salesforce states that searching CRM data and Slack carries no extra fee for users on eligible Unmetered seats, and that their usage shows as zero consumption in the Digital Wallet.

Costs appear in two places:

  • Data 360 objects. Answers that pull from connected Data 360 objects consume Flex Credits or Data Services Credits.
  • Users without an eligible seat. Anyone who loses an Unmetered seat moves back to normal Flex Credit or Data Services Credit billing.

So the budget question is not “what does Coworker cost” but “which users are on metered seats, and how much Data 360 do we plan to expose”. If you are still working out how Flex Credits are counted, start with our Flex Credits explainer, and read the Core, Advanced and Max editions change alongside it, since the new editions bundle credits differently.

Agentforce Coworker pricing: no extra fee for CRM and Slack search on eligible seats, credits for Data 360 answers and metered users
Who pays what for Agentforce Coworker.

Requirements to turn on Agentforce Coworker

If your org was not switched on automatically, you can turn it on yourself. The setup guide and the help FAQ list these requirements.

RequirementWhat you need
EditionEnterprise, Unlimited or Agentforce 1
Usage based licensingData Cloud or Data 360, a Foundations or Flex Foundations licence, and Flex Credits
Seat based licensingData Cloud or Data 360, plus Agentforce 1 Edition or Agentforce for Sales, Service and Industries
Admin accessSalesforce admin role, the Agentforce Coworker Admin permission set and the AI Search Setup permission set licence
User accessThe Access_Ai_Search permission set group

How to turn on Agentforce Coworker, step by step

  1. In Setup, type Agentforce Coworker in the Quick Find box.
  2. Select Get Started with Agentforce Coworker, click Turn On and confirm. Einstein turns on automatically at this point. Setup takes a few minutes, and Salesforce shows a warning if any prerequisite is missing.
  3. If your org has more than one data space, choose the data space, then select the data sources Coworker should use.
  4. In the Manage Users section, click Manage and assign the Access_Ai_Search permission set group to the users who should have it.
  5. For seat based access, open each user in Setup, go to Permission Set License Assignments, tick Unmetered User-Based AI and save.
  6. Test with two or three users from different roles before you widen access. Ask each the same question and check that the answers differ in the way their permissions say they should.
How to turn on Agentforce Coworker in six steps: Quick Find, Turn On, data sources, assign users, seat licence, test
The six steps to turn on Agentforce Coworker.

Troubleshooting common Agentforce Coworker errors

These are the errors Salesforce lists in its troubleshooting FAQ, with the fix for each.

ErrorLikely causeFix
Coworker does not appear, or “Not Enabled”User is missing the licence or the Access_Ai_Search groupCheck the permission set licence and group assignment for that user
401 UnauthorizedConnected App or OrgJWT set up incorrectlyConfirm the Connected App definitions and OrgJWT certificates match
400 Invalid IDAgent ID is truncated or in the wrong formatCopy the full 18 character Agent ID from the setup URL
500 Server Error or engine lookup failureBackend or model gateway interruptionHard refresh the browser, then turn Coworker off and on again
Invalid TokenThe session failed OAuth or JWT validationCheck client credentials and the gateway network path

Limits worth knowing before rollout

  • 30 language model calls per minute per user. Salesforce notes this is not technically enforced yet, but one question can trigger several calls, so heavy users will reach it sooner than the number suggests.
  • Record updates are CRM only. Coworker can read Slack and Data 360, but it can only write to Salesforce records.
  • Files over 75 MB are not indexed, so very large documents will never show up in answers.
  • English only during the beta. Salesforce says other languages follow at general availability.
  • Not yet available in GovCloud. It does work for authenticated Experience Cloud users.
  • Answers are only as good as the data. Duplicate accounts, empty close dates and stale opportunity stages all show up in the answer. An AI summary of messy data is still messy data, just faster.

A sensible Agentforce Coworker rollout plan for admins

Before rollout, it helps to have your core Salesforce Sales Cloud workflows documented, because Coworker surfaces whatever your automation writes to records, good or bad.

Whether Coworker is already on or you are about to switch it on, the order of work is the same:

  1. Check the status. Open Setup and see whether Coworker is active and who has the Access_Ai_Search group.
  2. Review sharing. Look at organisation wide defaults, sharing rules and the objects with public read. Fix what you would not want a quick question to reveal.
  3. Clean the high use objects. Merge duplicate accounts and contacts and fill the fields reps rely on, such as stage, close date and next step.
  4. Decide on Slack and Data 360. Connect Slack if your teams live there. Connect Data 360 sources only where the value is clear, set governance policies for any source that is not permission aware, and remember those answers consume credits.
  5. Start with a pilot group. Pick one sales team and one service team, collect the questions they actually ask, and write short guidance on what Coworker is good at.
  6. Measure. Track adoption, the questions that return weak answers and credit consumption for the first month before you expand.
Agentforce Coworker rollout plan: check status, review sharing, clean data, connect Slack and Data 360, pilot group, measure
A six step Agentforce Coworker rollout plan for admins.

If you need help with the sharing review, the data clean up or the Data 360 connections, our Salesforce managed services and Data Cloud consulting teams work on exactly these steps.

Coworker rollout review

Coworker may already be switched on. Is your data ready for it?

Coworker answers from whatever your users can already see. We review permissions and data quality first, so it gives the right answers to the right people.

  • Check whether Coworker is enabled in your org today
  • Find profiles and sharing rules that expose more than they should
  • A simple rollout plan for your first team

Get a Coworker rollout review

Certified Salesforce partner · 30-minute call, no obligation

Frequently asked questions about Agentforce Coworker

Is Agentforce Coworker part of AIforce?

Yes. At Dreamforce 2026 Salesforce grouped Agentforce Coworker, Claudeforce and Slackforce under AIforce, its new interface layer for reaching Salesforce data from any screen.

Is Agentforce Coworker free?

Searching CRM data and Slack carries no extra fee for users on eligible Unmetered User Based AI seats. Answers that use Data 360 objects consume Flex Credits or Data Services Credits, and users without an eligible seat are billed on credits.

Was Agentforce Coworker turned on automatically?

Possibly. Salesforce began automatic enablement on 10 August 2026 for orgs with Unmetered User Based AI seats and Einstein active. HIPAA, health, government and multi org setups were excluded.

How do I opt out of automatic activation?

Use the opt out link in the Salesforce notification email, or go to Setup, Agentforce Coworker, and turn on Opt Out of Auto Activation.

Can Agentforce Coworker see data a user cannot see?

No. It runs with the permissions of the person using it, so it can only read and change what that user can already read and change.

Does Agentforce Coworker work in Slack and Microsoft Teams?

Yes. Salesforce lists Slack, Microsoft Teams, Claude, ChatGPT and mobile as surfaces alongside the Salesforce search bar.

Which editions support Agentforce Coworker?

Enterprise, Unlimited and Agentforce 1 editions, with Data Cloud or Data 360 in place.

Why can’t I see Agentforce Coworker in my org?

Usually the user is missing the required permission set licence or the Access_Ai_Search permission set group. Check both in Setup before looking for anything else.

The bottom line

Agentforce Coworker is the most visible part of AIforce because it lives where every user already clicks: the search bar. It costs little or nothing for most seat based orgs, and it may already be running in yours. The real work is not switching it on. It is making sure your permissions and data are in a state where a fast, confident answer is also a correct one. Start there, pilot small, and expand once the answers hold up.

Categories
Agentforce Salesforce

Agentforce Named Agents: What Casey, Paige, Carter, Hunter, Marshall, Piper and Fin Actually Do

For two years Agentforce was something you built. You picked a topic, wrote instructions, wired up actions, tested it, and hoped the thing behaved in production. At Dreamforce in September 2026 Salesforce changed the starting point: seven named agents that arrive already configured for a specific job, six of them generally available on day one.

That shift is bigger than the names suggest. A prebuilt agent is an opinion about how a job should be done, baked into software. It will fit some orgs closely and fight others. This guide goes through all seven, what each one is actually scoped to do, which cloud it needs, what is shipping today versus later, and where we would not use it.

The seven Agentforce named agents at a glance

AgentJob it doesWhere it livesStatus
CaseyCustomer service resolution across voice, SMS, WhatsApp and web chatService CloudGA now
PaigeIT and HR service requests through Slack and employee portalsEmployee ExperienceGA now
CarterProduct discovery, comparison and checkout help for shoppersCommerce CloudGA now
PiperInbound lead engagement and qualification across web and emailSales and MarketingGA now
FinComplex customer experience workflows spanning channelsCustomer OperationsGA now
MarshallBack office process orchestration with deterministic executionBack OfficeGA now
HunterOutbound pipeline, from research through to outreachSales CloudPilot now, GA expected November 2026

1. Casey: the service agent that finally covers voice

Casey handles customer service issues end to end across voice, SMS, WhatsApp and web chat. The channel spread is the part worth noticing. Most deflection projects we see start and stop at web chat, because chat is where the transcript is clean and the integration is easy. Voice is where the volume and the cost actually sit, and voice is where a handover to a human has always been ugly.

What Casey is good at: repeat, policy driven questions where the answer exists in a knowledge article or a record. Order status, appointment changes, password and access issues, returns inside policy.

What it does not remove: the need for a clean knowledge base. An agent that reads bad articles gives confident bad answers faster than a human would. If your knowledge base has not been audited in a year, that audit is the first project, not the agent.

2. Paige: employee service, where the ROI case is easiest

Paige answers IT and HR service requests inside Slack and employee portals. Of the seven, this is the one with the least risk attached, because the audience is internal. A wrong answer to an employee costs a follow up ticket. A wrong answer to a customer costs the customer.

That makes Paige a sensible first deployment even for an organisation whose real target is customer facing automation. You get the operating experience, the prompt tuning habits and the escalation design without putting revenue on the line while you learn.

3. Carter: commerce assistance, not a storefront replacement

Carter works inside Commerce Cloud on product discovery, comparison and checkout assistance. It is the agent most often confused with something it is not. Carter does not replace your storefront, your catalogue structure or your merchandising rules. It sits on top of them, and it inherits their quality.

If product data is thin, if attributes are inconsistent between categories, or if the same item appears three times under different SKUs, Carter surfaces those problems to shoppers rather than hiding them. Catalogue hygiene is the prerequisite. We cover how that data flows in our guide to Salesforce ecommerce integrations.

4. Piper: inbound lead qualification

Piper engages and qualifies inbound leads across websites and email. The job is speed to lead and consistency of questioning, two things humans are reliably bad at on a Friday afternoon.

The caution here is qualification criteria. Piper applies whatever definition of a qualified lead you give it, at scale, without the quiet human correction that normally happens in the background. If your MQL definition has drifted from what sales will actually accept, an agent will expose that gap within a week. Fix the definition before you automate it.

5. Fin: the cross channel workflow agent

Fin is scoped to complex customer experience workflows that span channels, which is the vaguest brief of the seven and also the most interesting. Think of the cases that start in chat, move to email, require a back office action and end with an outbound confirmation. Those are the ones that fall through today, because no single team owns them end to end.

Fin is worth evaluating if your complaint volume is driven by process breakage rather than by product problems. It is not worth evaluating if your handoffs are undocumented, because there is nothing for it to follow.

6. Marshall: deterministic back office execution

Marshall orchestrates back office processes with what Salesforce calls deterministic execution. That word matters. Most of the anxiety around agents in finance and operations comes from non determinism: the same input producing a different sequence of steps on different days. Marshall is positioned as the answer to that, running defined processes in a defined order rather than reasoning its way to an outcome.

Where this fits: invoice handling, order processing, fulfilment exceptions, reconciliation steps that today sit in a spreadsheet and a person’s head. Where it does not: anything where the process itself is undefined. Marshall executes a process, it does not design one for you.

7. Hunter: outbound pipeline, and the one that is not GA yet

Hunter covers outbound sales pipeline from account research through to outreach, and it is the only one of the seven still in pilot, with general availability expected in November 2026. It is also the agent that debuts long horizon runtime, which is the genuinely new platform capability in this release.

Long horizon runtime means an agent can pursue a goal across days and weeks rather than inside a single session. It preserves memory between runs, survives interruption through durable execution, and can be steered mid flight without being restarted. For outbound that is the difference between a sequence and a campaign that adapts.

It is also the capability that deserves the most governance attention. An agent working a goal for three weeks accumulates three weeks of decisions nobody reviewed individually. Decide in advance what it is allowed to send, to whom, and what it must escalate.

What shipped alongside the agents

Three platform pieces came with the agent portfolio and change how the seven are used together.

Multi agent orchestration is generally available now, and routes work across the specialised agents so they operate as a team rather than seven separate deployments. This is what stops Casey and Fin from both claiming the same case.

AI Skills in Agentforce Coworker lets an employee teach a task once and have it reused across the workforce. It is in pilot now with general availability expected in October 2026.

Agent Optimizer covers building, refining, testing and analysing agent performance across the lifecycle, with general availability expected in October 2026. Until it lands, measurement of agent quality is still something you assemble yourself.

All of this sits inside the wider announcement Salesforce made at Dreamforce. If you want the commercial side, including which Flex Credit allowance comes with which tier, start with what AIforce actually changed and the Core, Advanced and Max editions.

When a named agent is the wrong answer

Four situations where we would tell a client to wait.

The underlying data is not trustworthy. Every one of these agents reads your records and your content. Duplicate accounts, stale knowledge articles and half migrated product data do not get smoothed over by an agent, they get amplified and put in front of a customer.

The process is undocumented. Marshall and Fin in particular assume a process exists. If the answer to “what happens next” depends on which person picks it up, an agent has nothing to follow.

Your volume is low. An agent earns its keep on repetition. A team handling forty cases a week will spend more effort tuning than it saves.

The job is judgement, not throughput. Complaint escalations that involve goodwill decisions, a renewal conversation with an unhappy account, anything with a legal or clinical dimension. Those stay human, with the agent doing preparation rather than resolution.

How to choose which one to turn on first

The order we recommend, and the reasoning behind it.

Start with Paige if you have an internal service desk. Lowest blast radius, fastest feedback loop, and the operating habits transfer directly to the customer facing agents afterwards.

Start with Casey if service cost is the board level number. It is the only one of the seven that touches voice, which is usually where the cost sits.

Start with Piper if speed to lead is the bottleneck. Measure current first response time before you deploy, because this is the one agent whose effect shows up in a single clean metric.

Wait on Hunter unless you are comfortable in a pilot. General availability is expected in November 2026, and long horizon runtime is new enough that early deployments are learning exercises as much as productivity ones.

Whichever you pick, scope one job, instrument it, and run it for a month before adding a second. Teams that switch on four agents at once cannot tell which one caused the change in their numbers.

What this means for existing Agentforce builds

If you already built custom agents, the named agents do not obsolete them, but they should prompt a review. Anything you built that maps closely to one of the seven jobs is now maintenance you may not need to carry. Anything you built that is specific to your business, your products or your regulatory position is exactly what should stay custom.

The practical exercise is to list your current agent topics against the seven and mark each as replace, keep or merge. Our Agentforce readiness checklist covers the questions to ask before either path, and the implementation partner guide covers how the delivery work is usually split.

The named agents are not the only way users reach Agentforce. Agentforce Coworker in the Lightning search bar routes a question to the right specialised agent on its own, and our Agentforce Coworker guide explains how that works and what it costs.

The named agents arrive alongside a busy platform release. See the Salesforce Winter ’27 release guide for the 15 changes to test in your org, including the new Agent Skills registry.

Frequently asked questions

Are all seven Agentforce named agents available now?

Six are generally available: Casey, Paige, Carter, Piper, Fin and Marshall. Hunter is in pilot, with general availability expected in November 2026.

Do the named agents replace custom built Agentforce agents?

No. They cover common, well defined jobs. Anything specific to your products, your industry rules or your internal processes still needs to be built, and the two run alongside each other under multi agent orchestration.

Which licences do the named agents need?

Each sits inside a product area, so Casey needs Service Cloud, Carter needs Commerce Cloud and so on. Consumption is drawn from the Flex Credit allowance that comes with your edition, which means two organisations on the same edition can see very different usage. Confirm current entitlements with Salesforce or your account executive before budgeting.

What is long horizon runtime?

It lets an agent pursue a goal across days and weeks instead of within one session, by preserving memory between runs, surviving interruptions and accepting steering mid task. Hunter is the first agent to use it.

Where do we start if we have never deployed an agent?

With an internal use case, a documented process and a single measurable metric. Paige on an IT or HR service desk is the usual low risk starting point.

Talk to us about which agent fits

Ashapura Softech is a certified Salesforce partner working with teams in the United States on Service Cloud, Sales Cloud, Commerce Cloud and Agentforce delivery. If you want a view on which of the seven fits your org, and what would have to be cleaned up first, get in touch.

Source: 21 Things We Announced at Dreamforce 2026 (Salesforce).