I run executive briefing programs, and I build the tooling that runs them.
My background is go-to-market strategy, enterprise customer programs, and AI-driven transformation work. Most recently I led an 8-figure flagship executive briefing program at Cisco. I don't come from an engineering background. I build things anyway, because the fastest way to find out whether AI actually helps a program is to make it help one.
This account is where the working software lives. The rest of the story is at andrewhahm.com.
ebc-ai-deployment-kit Where to apply AI in an executive briefing program, and where not to. A readiness assessment a briefing team can run on themselves in five minutes, the reference architecture for the automated half, and a worked spec for one automated step including the failure modes it has to be built against. The argument underneath it: the assembly is automatable, the judgment is the job.
claude-usage-monitor A Chrome extension that surfaces Claude plan limits, credit balances, and burn-rate projections in the browser toolbar. Built because there's no usage API for consumer plans and I kept hitting the ceiling mid-conversation. It reads undocumented JSON endpoints and renders limit buckets dynamically, so it survives schema changes instead of breaking on them. MIT licensed.
andrewhahm-site My site. One self-contained HTML file, no framework, no build step, with a Google Apps Script backend for the inquiry form. Deployed on GitHub Pages.
How much of an executive briefing operation should actually be automated, and which parts have to stay human. My working answer: standardize the content, customize the conversation. The prep, the capture, and the follow-through are automatable. The room is not.