How is your team collaborating while working with coding agents?

From what I have seen, the teams that keep things simple tend to work better with coding agents. Most collaboration still happens in GitHub issues/PRs, and the agents are just another tool in the workflow. When too many docs and tools get involved, things start to feel messy quickly.

Right now everyone is trying to figure out what LLMs mean for individual productivity. We still pretend we are doing most of the stuff when it comes to team dynamics and behave as individuals.

I’m curious to see what comes next and how team dynamics and collaborative work will change once we are past that phase.

In my vision people will answer questions that other humans or LLMs have based on tons of documents and accumulated context that is shared and things will happen.

What is missing yet is how to share all that context. I don’t think any one current solution will be the winner because they are all too brittle and crumbly. It’s going to be something new and everyone from the company will have to use it to be effective.

There’s probably loads of opportunity here.

We are using https://mfbt.ai/