Salesforce has shipped Agentforce directly into Slack as a first-class chat participant. Users can DM an agent or @-mention one in a channel the same way they'd tag a coworker, and the agent responds inline — pulling Salesforce data, updating CRM records, generating customer briefs, or posting a deal recap without anyone having to leave the conversation.
The interaction model is the whole point. Agentforce-in-Slack discards the 'launch a separate AI app' pattern and treats agents as teammates that live in the same threads as human colleagues. A sales rep can tag an agent in a thread, ask it to pull the latest opportunity data, and get a researched answer grounded in Data Cloud — then forward the response to a teammate with the record already summarized. The work stays in Slack. The context stays with the conversation.
Three agent templates ship out of the box. The Customer Insights Agent uses Salesforce data to summarize customers, find internal subject matter experts, and generate customer briefs. The Employee Help Agent resolves internal questions using org knowledge. The Onboarding Agent generates onboarding canvases and answers common new-hire questions from a custom knowledge base. Organizations can also build custom agents that plug into the same Slack surface.
The strategic read is clear. Slack is betting that the AI-at-work winner isn't a new standalone interface — it's the chat surface companies already live in, with agents running alongside the humans. That puts Salesforce on a direct collision course with the Copilot-Agents-in-Teams stack and with OpenAI's new Workspace Agents, which also ship with Slack integration. The 'where do agents live' question is now being settled across three simultaneous product launches.