Skip to content
News TechCrunch Jul 2026

TechCrunch: Jack Dorsey launches Buzz, a team chat built for AI agents

Jack Dorsey announced Buzz on July 21, 2026 — a workplace group chat platform built by Block, the company behind Square, Cash App, and Afterpay. Unlike Slack, which added AI features to an existing collaboration tool, Buzz was designed from the start to include AI agents as participants in conversations rather than treating them as bolt-on integrations.

Dorsey describes the platform as “model-agnostic, decentralized, self-sovereign, and open source.” The open-source nature means teams can inspect the codebase, deploy their own instances, and extend the platform without depending on Block’s roadmap. In practical terms, Buzz consolidates group chat and GitHub-style project management into a single interface where human contributors and AI agents share the same thread context.

The timing reflects a real friction point that product teams have started running into: as organizations deploy multiple AI agents for tasks like drafting, research, code review, and customer support, those agents often work in isolation from the tools where human decisions actually happen. The result is a fragmented context — agents process information in one system while humans coordinate in another. Buzz bets that the right architecture is a unified workspace where agents and people operate in the same conversation stream.

For product managers, the immediate implication is less about switching away from Slack and more about what Buzz signals for the near term: enterprise collaboration software will likely need to treat agents as first-class participants rather than API add-ons. Teams evaluating workflow tooling in late 2026 should expect this expectation to shape vendor roadmaps across the category.

Buzz remains in early stages and Dorsey’s track record with late-stage product execution is mixed, so teams with established Slack infrastructure have little reason to migrate now. For early-stage startups building AI-heavy workflows from scratch, it may be worth evaluating.