Jack Dorsey is taking on Slack with Buzz, a group chat platform for teams and their AI agents

We have reached a point in the software cycle where every legacy tool feels like it is built for a world that no longer exists. Slack was intended to kill email, but instead, it just gave us more noise to manage. Now, Jack Dorsey is stepping back into the ring with Buzz. It is a group chat platform, yes, but the focus isn’t just on humans yelling at each other in channels. It is designed to host the new workforce: a mix of people and autonomous AI agents.

The Workflow Problem

For the last decade, we have treated AI in the workplace as a plugin or a sidecar. You have your chat in one window, and you have your LLM interface in another. If you want to integrate them, you usually have to mess with complicated API bridges or third-party bots that feel clunky and limited. The current infrastructure of work is human-centric, which sounds nice, but it ignores how work is actually getting done in 2026.

Dorsey’s play with Buzz seems to be based on a simple observation: if we are all going to have personal AI agents doing our heavy lifting, those agents need a place to live. They shouldn't be hidden behind a command line or a separate dashboard. They need to be in the room where it happens, interacting with other agents and humans in a streamlined interface.

The Core Proposition of Buzz

Buzz is being positioned as a direct competitor to Slack, but the architecture is fundamentally different. In a traditional team chat, an AI bot is an intruder. In Buzz, the AI agent is a member. This isn't just a semantic difference; it changes how data flows through a corporation. When an agent is a first-class citizen in a chat, it can maintain context, track deliverables, and communicate with other agents without a human acting as the middleman.

From building a product perspective, this solves one of the biggest bottlenecks in current workflows: context switching. We spend a massive amount of time explaining things to our tools. If a tool is consistently present in the conversation, the overhead of "bringing the AI up to speed" disappears. It is always up to speed.

What This Means for Technical Founders

If you are building in the crypto or AI space right now, you need to watch this closely. We are moving away from the "app" era and into the "agent" era. Founders shouldn't just be building features; they should be building autonomous entities that can inhabit platforms like Buzz. The interface becomes less important than the utility and the communication protocol of the agent itself.

There is also a significant intersection here with decentralized identity. If agents are participating in team chats and making decisions, we need to know who owns them, who authorized them, and how they are being paid. I suspect we will see a heavy lean toward protocols that allow these agents to operate with their own wallets and cryptographic signatures. You can't have a truly autonomous agent if it still has to ask a human to swipe a corporate credit card every time it needs to pull a dataset.

The Skeptic's View

As much as I like the vision, we have to talk about the reality of Jack's track record. He is a visionary, no doubt, but he also tends to lean into ideological purity over user experience. If Buzz is too decentralized or too complex for the average HR manager to understand, it will suffer the same fate as many other well-intentioned tools: it will become a playground for developers that the rest of the world ignores.

Slack’s moat isn't its technology; it’s the fact that everyone's company data is already there. Moving a 500-person team to a new chat platform is a nightmare. To win, Buzz doesn't just need to be "better with AI." It needs to prove that the friction of moving is worth the gain in productivity. If the AI agents in Buzz can actually save a team 10 hours a week by automating administrative rot, then people will switch. If it’s just Slack with a fancy new coat of paint and some GPT integration, it’s DOA.

Why the Agent-First Approach Matters

Most people are still thinking about AI as a search engine. They use it to look things up or write emails. But the real value is in agency—the ability for software to take an action on your behalf. In a typical Slack channel, if I ask a bot to "book a flight," it usually spits out a link. In an agent-first environment like Buzz, my agent should talk to the travel agent's bot, negotiate the price, and finalize the booking within the thread.

This represents a shift from communication to execution. We have spent the last twenty years improving how we talk about work. We are finally starting to build tools that actually do the work. For builders, the opportunity isn't in building the chat platform; it's in building the agents that will populate it.

Takeaway for Builders

• Stop building standalone dashboards; start building agents that can live in shared environments.
• Focus on context retention. The value of an agent is directly proportional to how much it remembers about a project.
• Prepare for an economy where software talks to software more than humans talk to humans.

We are entering a phase where the "team" is no longer just a group of people on a payroll. It is a hybrid network of humans and autonomous code. Buzz is the first major attempt to give that network a home. Whether it succeeds or fails, it confirms that the way we organized work in the 2010s is officially dead. If you’re still building for at a 2015 workflow, you’re already behind.

Read the original at TechCrunch AI →