How to run a team of Grok bots like a project team

Ever had your Grok Bot chats turn into a mess? A post by Eric Zakariasson on X shares a clever fix: run them the way you'd run a human project team. Instead of one long chat per bot type, you create one channel per project and staff it with the right bots.

Why separate bot chats fall apart

Most people start with a Coder chat, a Writer chat and a Researcher chat. That's fine until you're juggling several projects at once. Then context gets mixed, updates bleed into each other, and it's hard to tell signal from noise. The bot doesn't know which project you mean unless you keep repeating it.

Borrowing the human team pattern

Human teams solve this with projects. Every project gets a spec, a plan and a channel, and tasks inside it go to different people. Zakariasson copied that pattern almost exactly.

He set up two databases in Notion: Projects and Tasks. Then he built a Projects Manager bot, or PM bot, with a /Project Ops skill. The PM takes care of the meta work: create a project, open a channel and staff it with the right bots.

One project, one channel, one roster of bots

Each project becomes one channel in the sidebar. Inside it you've got the PM bot, the specialist bots and the user. The roster is visible right there in the channel, so everyone knows who's on the project.

A few staffing rules stop the system from getting bloated:

Reuse the bots you have first. Coder, Researcher, Writer or anything else already on the bench gets priority.

Propose at most five bots besides the PM. Each channel tops out at six bots total.

Only create a new bot when nothing on the bench fits, and only once the user approves it. The new bot can be a specialist you reuse later, or a temporary one that goes away with the project.

How the work really moves

Once the channel exists, the user scopes tasks with the bot roster. The PM keeps an eye on progress and keeps the Notion databases updated. If a bot gets stuck or needs input, it marks its task as Blocked and pings the user in the channel.

So the user can often just watch cards move, without stepping in at every turn. And here's the fun twist: the bigger the system gets, the more it looks like something built for humans. A board, a manager, specialists claiming tasks, a blocked column and a channel.

Why this pattern is worth paying attention to

It's still experimental, but the idea underneath is practical. AI assistants are getting good enough that the model isn't the bottleneck anymore. Coordination is. Once you use bots for real work, treating them as a team instead of separate chat windows is a natural next step.

Try it yourself

Choose one real project that has several moving parts.

Set up a Projects database and a Tasks database somewhere you can update easily, like Notion.

Define a small bench of specialist bots you already use, like Coder, Writer and Researcher.

Make one PM bot. Its only job is to create the project, open the channel and staff it from the bench.

Cap the channel size so it doesn't turn into a wall of noise.

Let bots mark tasks as blocked and ask for help, rather than guess.

What to take away from this

The best AI workflows will probably look a lot like the best human ones. Clear project scope, a visible roster, assigned tasks and an explicit handoff when something's stuck. Bots don't need a new management theory. They need the old one, used consistently.

Source: a post by @ericzakariasson on X, shared on 27 August 2026.