Salesforce is giving Slack a broad AI-focused overhaul, with 30 new features announced at a small gathering in San Francisco on Tuesday. The update builds on Salesforce’s wider effort to remake its business around AI, and it places Slackbot at the center of a more automated version of Slack.
The changes are designed to move Slack beyond enterprise messaging. Salesforce is positioning Slack as a place where business work can be planned, routed, summarized and followed up, with AI handling more of the connective work between people, channels, apps and data sources.
Slackbot gets a larger role
The most important part of the announcement is a major upgrade to Slackbot, Slack’s AI agent. Salesforce had already updated Slackbot in January with agentic capabilities, including the ability to draft emails, schedule meetings and search through an inbox for specific information.
The new set of features expands that direction. Rather than acting only inside a chat workflow, Slackbot is being framed as an agent that can help coordinate work across multiple business contexts. That includes Slack channels, connected apps, data sources and outside tools.
The 30 new features will be available in the coming months. Salesforce presented them as part of a larger push to make Slack more useful for core business processes, not just conversation.
Reusable AI-skills are the central feature
One of the most notable additions is what Salesforce calls reusable AI-skills. These let users define specific tasks for Slackbot and then apply those tasks across different situations.
Salesforce says Slackbot includes a built-in library of AI-skills. Users can also create custom versions for their own work. The idea is that once a skill exists, an employee does not have to rebuild the same process every time a similar need comes up.
For example, a user could trigger a skill with a simple Slack command such as “create a budget” for an upcoming event. Slackbot would then gather relevant information from company Slack channels, connected apps or data sources, and use that information to create an actionable plan.
The workflow can continue beyond the plan itself. In that example, Slackbot would also automatically set up a meeting to discuss the plan and invite relevant employees based on their titles.
That matters because it shows how Salesforce wants Slackbot to operate: not only as a helper that answers questions, but as a system that can assemble information, turn it into next steps and coordinate the people needed to move those steps forward.
Slackbot connects to outside agents and tools
Slackbot now also functions as an MCP (Model Context Protocol) client. In practical terms, that means it can connect to and coordinate with outside services and tools.
One of those tools is Agentforce, Salesforce’s AI agent development platform launched in 2024. Through that connection, Salesforce says Slackbot can “route work or prompt questions to Agentforce or any agent or app in your enterprise”.
The company says the agent can find the most relevant and efficient path for information without human intervention. That is a meaningful shift in how Slack is being described. Instead of being only the place where employees discuss work, Slack becomes a place where AI can decide which system or agent should handle a task.
This also explains why Salesforce is emphasizing Slackbot’s ability to work across different services. The more systems Slackbot can reach, the more Slack can become a central layer for enterprise AI workflows.
Meetings, recaps and desktop context
According to Rob Seaman, EVP & GM of Slack, Slackbot can now transcribe meetings and summarize them. If someone misses an important detail during a meeting, they can ask Slackbot for a recap that includes action items assigned to them.
That feature fits the broader theme of the update: Slackbot is being designed to reduce the amount of manual follow-up required after work conversations happen. Meeting notes, assigned tasks and next steps can be brought back into the Slack workflow.
The agent can also operate outside Slack and monitor desktop activities. Salesforce lists “your deals, your conversations, your calendar, and your habits” as examples of the data it can use.
Based on that wider context, Slackbot can make actionable suggestions or draft follow-ups for critical tasks. Seaman has said privacy protections are built into the design and that users can adjust permissions as needed.
That permission layer is an important part of the product framing. Salesforce is describing a more context-aware Slackbot, but also noting that users have controls over what the agent can access.
Salesforce wants Slack to become core infrastructure
The strategy behind the announcement is clear: Salesforce wants Slack to be more than an enterprise communication tool. The company is trying to turn Slack into a more versatile platform for handling business tasks.
The AI-heavy update gives Slackbot more ways to act on information, connect with outside systems and support everyday work. Reusable AI-skills, MCP support, Agentforce integration, meeting summaries and desktop context all point toward the same goal.
Marc Benioff let his team explain the major features on Tuesday, but he also reflected on Salesforce’s ownership of Slack. During his keynote, he said the five years since Salesforce acquired Slack had been an “incredible jou rney,” one that had de livered “two and a half times revenue growth.”
He added: “We have about a million businesses running on Slack. It’s been a huge growth story.”
For Salesforce, the next phase appears to be making those businesses rely on Slack for more than messages. With these 30 AI features, Slackbot is being pushed into planning, routing, summarizing and follow-up work. The result is a version of Slack built around AI agents as much as human conversation.