AI at KODE

We're an AI-forward company. Everyone uses AI throughout their work, and people stay accountable for what we ship.

How we use it

AI is part of how everyone works every day, not only engineers. Product uses it to write briefs and build designs and prototypes, and security, support and the rest of the team use it in their own work. Engineers use it from the first story to the release plan.

Today we use tools such as Claude, Codex and Amp, and the list changes as the tools improve. We'd rather try a promising new tool than hold on to an old one.

We learn it together

AI changes every month, so nobody figures it out alone. We share what we discover as we go, in Slack, in demos and at our engineering roundup, and we keep refining our best practices as a team. When someone finds a better way to work, it becomes everyone's way to work.

People stay accountable

• The engineer who opens a pull request owns everything in it, however it was written.
• Before a pull request, a second AI agent reviews the code, and then a person reviews it too.
• AI review adds to human judgment. It never replaces it.

Safety comes first

We use AI only in ways that meet our requirements for protected health information (PHI), security and training. Healthcare data deserves that care, and it isn't negotiable.

What we look for

At every level, we expect engineers to use AI well: to pick the right tool, check the output, and spot its typical mistakes. As engineers grow, they shape how the team uses AI. See the engineering levels.