AI Productivity Is About More Than Writing Code Faster

Increasing developer productivity with AI isn’t just about writing code faster.

The conversation around AI coding tools like Claude Code et al. tends to center on one thing: writing code faster. And that’s a real win.

But I’ve been thinking about a different bottleneck.

Before developers can tell these tools what to work on, there’s a tax: reconstructing context. What did we decide in that Zoom call? What’s the current state of this epic? Why does the Jira contradict that Slack thread from three weeks ago?

The more interesting application of these models, to me, is using them to continuously ingest and distill the information your team already produces — chat, email, tickets, wikis, meeting transcripts, code review threads — so that context is ambient rather than something each developer has to manually reconstruct each morning. Plus, these tools enable each developer to customize how all of this gets put together, instead of relying on a one-size-fits-all solution handed to them.

AI-assisted coding gets the headline. But AI-assisted thinking — reducing the archaeology so developers can spend more time on the actual hard problems — that might be the more durable productivity gain.