01 /
Claude Code, Codex or open-source models: how we pick the tool for the job
Closed or open-source? There's no best model, only the right one for the job, the budget and the data. Here's how we choose.
Notes
Notes from the build.
What we've learned building AI systems for real teams. Short, practical, no hype.
02 /
Build or buy: when a custom AI system beats another software subscription
Off-the-shelf tools are great until your work doesn't fit them. A simple way to tell which side you're on.
03 /
What “you own it” actually means: your accounts, your code, your prompts
Ownership is a checklist, not a slogan. Here's exactly what should be in your name when a build is done.
04 /
The five manual jobs every deal team should hand to AI first
Not the flashy stuff. The repetitive work that eats a deal team's week, and the order we'd hand it over in.
05 /
Is this process worth automating? A 10-minute test
Five questions, ten minutes, one honest answer. Run it before you spend a dollar on AI.