Salesforce has announced more than 30 new Slackbot capabilities, repositioning Slack as what it calls “the new interface for work,” a single environment where employees, AI agents, data, and enterprise apps operate together.
At the heart of the update is a push to put AI-powered customer service and employee support in the same place where work already happens, eliminating the context-switching that has become one of the biggest drags on contact center productivity.
For customer service operations used to running at scale, the problem Salesforce is attempting to solve is a familiar one.
Companies are now deploying an average of 20 AI agents per year, but agents scattered across disconnected platforms create their own burden. Employees don’t know which agent handles which task, context gets lost between tool switches, and the efficiency gains AI promised quietly evaporate.
Speaking during a pre-briefing ahead of the announcement, Rob Seaman, EVP and GM of Slack, said:
“I shouldn’t have to know which agent’s a specialist in which task. You should be able to turn to almost a concierge personal agent.”
The goal, as he framed it, is “don’t make me think.”
That concierge is now Slackbot.
What’s Actually Launching
The 30+ new capabilities touch almost every part of the working day.
Meeting transcription captures discussions in real time, summarizes decisions, and triggers follow-ups the moment a call ends.
A desktop integration lets Slackbot read context across enterprise apps and take action using the employee’s existing permissions – drafting an email, updating a CRM record, or pulling a report – without switching applications.
Reusable AI Skills let teams define a task once and have Slackbot execute it automatically.
Deep research extends its reach across Slack history, Google Drive, OneDrive, and the web. An expanded MCP client connects it to Agentforce, third-party tools, and more than 6,000 apps across the Salesforce ecosystem, making Slackbot a routing layer for the entire agent network.
There’s also a voice interface and a memory capability that retains user preferences across sessions.
On security, Slackbot operates only within the authenticated user’s existing permissions, with no super-user access and explicit opt-in required before acting in external systems.
A new confidential channels feature lets teams mark conversations as AI-restricted, giving compliance-sensitive organizations a clear control mechanism.
The Numbers
Salesforce has been running this internally since January. The company reports 55,000 weekly active users and 87% week-over-week retention across its own workforce, with Seaman noting Salesforce is now tracking to one million weekly Slackbot users.
Business value studies that previously took weeks now take minutes.
Elsewhere, Engine, a travel and spend management platform serving more than one million travellers across 30,000 businesses, built its first AI agent on the platform in 12 days.
That agent (EVA, or Engine Virtual Agent) now autonomously resolves more than 50% of travel cases without any human intervention.
Where the CX Story Sharpens
For contact center leaders, the Engine deployment is the most instructive case study.
Beyond EVA’s 50%+ autonomous resolution rate, Elia Wallen, Engine’s Founder and CEO, described a much broader transformation.
New hire onboarding is handled through Slackbot, drawing from the company’s internal knowledge base.
Sales reps surface competitive intelligence in seconds during live client calls.
And perhaps most striking for anyone running a QA function, Engine now runs sentiment analysis and quality checks on 100% of customer calls. Previously, manual sampling covered roughly 1%.
In detailing why the solution works, Wallen explained that “context is where the hundred IQ points are.”
“You have all the context, you can answer anything.”
An agent with the full history of a customer relationship, a conversation thread, and relevant business data is categorically more useful than one operating on a partial picture.
reMarkable echoed the theme. Peter Stoltz, VP and Head of CIO Office, said his teams use Slackbot’s natural language search to pull context from across DMs and channels –summarizing weeks of conversation or confirming specific agreement terms in seconds.
“Slackbot consistently delivers the quick, accurate answers we need to stay aligned,” he said, “saving our teams hours of manual work each week.”
The Broader Signal
The companies seeing the most tangible returns from AI aren’t necessarily running more sophisticated models than their peers.
Instead, they’re the organizations that are reducing the friction between employees and agents to near zero.
When AI lives in the same interface where work already happens, adoption stops being a change management problem. As Seaman put it:
“The companies that have the unfair advantage are the people that are using AI and agents in the most impactful way.”
