TLDR: Slack, a Salesforce company, has announced a significant evolution of its platform, opening it up for advanced AI agent and application integration. Key introductions include a Real-time Search (RTS) API for instant access to conversational data, a Model Context Protocol (MCP) server to standardize LLM connections, and enhanced developer tools. These updates aim to empower businesses to build context-aware AI agents that can securely interact with Slack’s rich conversational data, streamlining workflows and boosting productivity.
Slack, the popular collaboration platform by Salesforce Inc., is undergoing a major transformation, positioning itself as an open platform for artificial intelligence agents, applications, and deep integration with conversational data. Announced on October 1, 2025, these updates are designed to enable developers to securely connect AI tools to the everyday workflows of employees and customers, leveraging Slack’s extensive repository of organizational knowledge.
At the heart of this evolution are several new developer tools. A crucial component is the Real-time Search (RTS) API, which provides on-demand, secure access to current discussions, files, and channels within Slack. This eliminates the need for bulk data transfers and local storage of sensitive information, allowing AI applications to retrieve only the necessary, up-to-date context. Kurtis Kemple, head of Slack developer relations at Salesforce, highlighted the efficiency gains, stating that this approach significantly reduces the time and energy developers spend writing complex code or making multiple API calls to achieve tasks.
Another pivotal introduction is Slack’s own Model Context Protocol (MCP) server. Developed and open-sourced by Anthropic PBC, the MCP server acts as a standardized communication layer between large language models (LLMs), AI applications, and agents, and Slack’s contextual information. This protocol simplifies integrations, reducing the manual effort required to define agent tasks or implement complex connectors for various LLMs. It ensures that AI agents can find information and act autonomously on behalf of Slack users, leading to more accurate, relevant, and personalized outcomes by reducing ‘hallucinations’ and increasing reliability.
These advancements are part of Salesforce’s broader ‘Agentforce’ initiative, an agentic layer deeply integrated with Data Cloud, the Einstein Trust Layer, and the Salesforce Platform. Agentforce in Slack aims to create a centralized digital home where humans and AI can collaborate seamlessly. Aaron Levie, CEO of Box, emphasized this vision, remarking, ‘You go to a single conversational interface in Slack, and you can interact with humans and AI. Agents are doing the work, and we’ve dreamt of that for decades. Now we’re finally seeing it.’
The benefits extend beyond technical simplification. By securely connecting AI agents to conversational data, Slack aims to unlock an organization’s collective intelligence, which was previously siloed. The RTS API, for instance, adheres to existing user access permissions, ensuring that AI applications only retrieve data relevant to their query, thereby maintaining data security and privacy.
Also Read:
- ServiceNow Unveils Unified Agentic AI Platform for Enterprise Transformation at Knowledge 2025
- Snowflake Unveils Advanced AI Agents, Promising Rapid Deployment and Enhanced Business Intelligence
Several prominent companies are already leveraging these new features to build advanced applications for Slack users, including Anthropic, Cognition Labs, Cursor, Dropbox, Google, Notion, Perplexity, Vercel, and Writer. This move signifies Slack’s ambition to evolve beyond a mere communication tool into an intelligent, agentic work operating system, serving as a command center for orchestrating tasks and interactions between human and AI workforces.
