Skip to content
TECH CEO Daily
SaaSAnalysis

Salesforce's Agentforce Coworker reached 100,000 activated users in 35 days as AI coworker tools multiply

Every major software vendor now wants to be the assistant employees open first. Without a plan, companies will pay for several that read the same email and files.

By · Editor

· 3 min read · Fact-checked

The 60-second brief

  • 1Salesforce said Agentforce Coworker reached 100,000 activated users within 35 days of launch.
  • 2Anthropic, Superhuman, Adobe and Meta each pushed AI coworker tools between September 10 and 28.
  • 3CIOs should choose a primary assistant per role, audit connector permissions and pay only for measured use.

The news

Salesforce (CRM) said on September 15 that Agentforce Coworker, its AI assistant, reached 100,000 activated users within 35 days. Between September 14 and 28, Anthropic, Superhuman and Meta (META) also expanded AI coworker tools, leaving CIOs with overlapping assistants competing for the same work.

Agentforce Coworker is now generally available inside Salesforce's Lightning interface. Salesforce Ben reported that it handles natural-language search and task routing, pulling from Salesforce, email, calendar and Drive, and can be used from Salesforce, desktop, mobile, Microsoft Teams, ChatGPT or Claude. Separately, Slack's chief marketing officer, Ryan Gavin, told SiliconANGLE that Slackbot has 1.5 million monthly active users since January, the fastest adoption of any Salesforce product.

Anthropic moved on September 16, merging its Claude chat and Cowork tools into one interface that routes requests automatically and adds tools for building documents and presentations, TechCrunch reported. The change is rolling out to Pro and Max subscribers first over the coming weeks, with free and Team plans to follow later. Anthropic is also Salesforce's partner for Salesforce in Claude, which puts CRM data inside Claude and is available to all Salesforce customers in beta, Salesforce said.

Superhuman, which says it serves more than 40 million people, bought AI notetaker Fathom on September 14 to feed meeting transcripts into its Go assistant. At Adobe (ADBE), CX Enterprise Coworker, an AI agent for marketing and customer engagement workflows, has drawn more than 1,700 customers and early adopters since it became generally available in June, Adobe said on its September 10 earnings call.

Then, on September 28, Meta launched Meta Enterprise Platform to sell its Muse assistant, a business agent and developer tools to companies, TechCrunch reported, with former MongoDB chief executive CJ Desai leading the effort. TechCrunch's report did not include pricing.

The numbers

Agentforce Coworker activated users
100,000 in 35 days
Slackbot monthly active users
1.5 million
Superhuman users (company figure)
More than 40 million
Adobe CX Enterprise Coworker customers and early adopters
More than 1,700

Why CEOs should care

CIOs should decide which assistant is the front door for each role before employees decide for them. A seller may live in Salesforce, a designer in Adobe and a knowledge worker in Claude or Microsoft Teams, but paying for three general-purpose assistants per person rarely makes sense. Set a primary assistant by role, treat others as features inside specific apps, and publish a short list of approved tools. Watch activation figures closely: activated users are not the same as weekly active users doing real work.

CISOs face a connector problem. Coworker tools gain value by connecting to email, calendars, file stores and chat, often through broad OAuth grants, the permissions users click to let one app access another. Each new assistant is another copy of access to sensitive content. Audit which assistants hold which scopes, remove unused grants, require admin approval for new connectors and confirm each vendor's data retention and model-training terms in writing.

CFOs should consolidate spend around measured use. Pricing varies across vendors, from per-seat add-ons to Salesforce's usage credits, which makes side-by-side comparison hard. Ask each vendor for usage reports by user, cancel seats that go unused for a quarter, and negotiate the right to swap seats between AI products within a vendor's portfolio.

The bigger picture

The contest is over the first screen employees open each morning. Whoever owns it controls which data the assistant sees and which workflows it triggers. Application vendors are fighting on two fronts: building their own coworkers, and opening their data to rival assistants on their own terms. Salesforce's AIforce layer lets Claude and Slack reach CRM data, while a September 17 help article says calls by third-party agents will consume paid Flex Credits. If other vendors follow Salesforce's lead, buyers could end up paying for access to the data that makes an assistant useful.

The sprawl has a security side as well. Each coworker that takes actions, rather than only answering questions, needs its own identity and permissions. Salesforce's September 17 requirement that agents calling its platform be registered with their own credentials is an early example of platforms forcing that discipline. Companies that have already narrowed their list of approved assistants will find those registration projects far simpler.

What’s next

Workday (WDAY) will show AI agents for HR and finance at its Rising conference, October 12 to 15 in Las Vegas. Anthropic plans to extend its merged interface to Team plans, and Meta's enterprise pricing was not reported at launch. Companies should set their assistant policy before those launches add another round of trials.

What “Fact-checked” means

Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.

What we checked
Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
How
A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
Who
The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, . A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
If something is wrong
“Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error

How we fact-check →

Companies in this story

SalesforceAnthropicSuperhumanMetaAI agents

Earlier coverage of Salesforce

All Salesforce coverage →

Written by

Editor · Technology & Business Writer

Hussein is a writer and business technology enthusiast focused on the intersection of technology, entrepreneurship, finance, artificial intelligence, and digital innovation.

CoversAICybersecurityBig TechSaaSStartupsFintech

About this story. Researched from primary sources whenever they are available and fact-checked before publication.

Published by Tech CEO Daily, an independent publication. Masthead · Editorial standards

Follow Tech CEO Daily on Facebook for the day’s top stories in your feed.

Free newsletters

The technology briefing for people running businesses.

Daily, weekly, bi-weekly or monthly. You choose.

How often

The Daily Brief · Monday to Saturday, 7 a.m. ET

Free forever. One click to unsubscribe. We never sell your email.