Slack Is Building the Agentic OS for Modern Enterprises

Once known primarily as a messaging app for teams, Slack is now emerging as the agentic operating system (OS) of the enterprise. Salesforce, which acquired Slack in 2021, is positioning the platform as a unified workspace where people, data, apps, and AI agents collaborate in real time to accelerate productivity.

“Slack has been a key component of the conversational layer for Salesforce for some time now,” said Himanshu Rajpal, regional sales director, Salesforce India, in an exclusive interview with AIM. 

The company said Slack will now serve as the conversational interface for Salesforce, with new integrations for Agentforce sales, IT and HR service, and Tableau. Teams can interact with CRM data and agents through natural language “right in the flow of work,” the company stated.

Slack’s updated Agentforce Sales lets sales teams manage customer records and pipelines directly in Slack without toggling between tools. Agentforce IT Service and HR Service bring instant AI support for common employee requests, while Agentforce Tableau embeds live dashboards into conversations.

“Teams, people, employees, stakeholders really are getting tired of switching between apps, changing context, and switching between different applications and tools all the time. They need to be able to drive quick decision-making through a singular interface,” said Rajpal. 

​​Moreover, developers can now build AI agents using Slack’s AI Developer Toolkit, with pre-built components and best practices for integrating large language models (LLMs).

What Agentic OS Means for Salesforce

Slack also introduced native AI capabilities within the platform, including a rebuilt Slackbot acting as a personal AI assistant and a Channel Expert agent that delivers instant answers from company knowledge. These features aim to help employees manage tasks, find information, and make faster decisions.

To support developers and partners, Slack has rolled out new Real-Time Search API and Model Context Protocol (MCP) server features. These tools allow companies like OpenAI, Anthropic, Google, Perplexity, Writer, Dropbox, Notion, and Vercel to build intelligent agents that run natively inside Slack.

As Slack expands AI integrations through MCP, Salesforce is doubling down on safety and control. Rajpal said Slack’s AI architecture is built on three pillars — trust, governance, and context. 

“Trust is about data security and auditability. Governance defines the rules of data access. Context ensures that every AI outcome is relevant to the user,” he said.

Productivity and Real-World Impact

Salesforce has already seen the benefits internally. Rajpal shared that 86% of its employees are now using Agentforce to get answers and take action directly within Slack. The company added that the tool is on track to save more than 500,000 employee hours annually.

Another example he shared was Anthropic, which uses Slack as its development hub and has cut product development cycles by 60%, saving an estimated $4.5 million annually. Rajpal said it’s still early in the journey, but the productivity gains so far have been significant.

Built on Agentforce

At the core of Slack’s transformation is Agentforce, Salesforce’s AI framework that powers the agentic layer across its ecosystem.

“Agentforce is deeply embedded within Slack,” Rajpal said. “When you talk about the agentic operating system for the modern workplace, it’s powered by Agentforce, whether it’s Slackbot, Channel Experts, or Service and Sales agents.”

This deep integration ensures that every Slack interaction can become an intelligent, contextualised action.

Moreover, Rajpal pointed to three key factors that set Slack apart from its competitors. The first is live permissioned conversational graphs, which provide real-time, contextual insight into customer interactions. 

The second is platform-level security, which ensures governance, observability, and trust across workflows. The third is actionable AI agents, which move beyond simple search and suggestions toward real, workflow-based execution.

“We’re now seeing AI agents turn conversations into workflows, and workflows into actionable outcomes and insights,” Rajpal said.

With over 2,800 connectors and 100,000 AI-powered apps already built on Slack, the platform is quickly becoming not just a workspace, but an intelligent operating system that learns, acts, and evolves alongside its users.