Bob Michaels/ai
All projectsCase study 04 of 10

Built to a seven-day deadline

Candidate AI Visibility Platform

Measures how five AI engines describe and rank political candidates, then writes campaign content only from verified findings.

Built to a seven-day deadline for a live webinar. It queries five AI engines, scores how each describes a candidate, and writes campaign content from the findings that passed verification.

When
May to Oct 2026
Status
Deployed, used in a live webinar
days to the first live use
7
commits
680
lines of code
103k
test functions
1,181

What this does not claim

Campaign outcomes are not measured here. The proof is that it shipped on deadline and ran live.

01

The problem

Voters now ask AI engines about candidates. Campaigns had no way to see what those engines say, and a wrong claim in campaign content is expensive.

02

What I built

A measurement system across five AI engines and a content writer that is allowed to use only findings that were verified. It runs as a deployed web app and API with background jobs and error monitoring.

03

What it produced

It was used in front of an audience in a live webinar on 2026-06-03, then kept growing through about 28 planned phases.

Measured from the repository

680 commits, about 103k lines, 1,181 test functions, 35 migrations, about 28 planned phases.

The stack

  • Python and FastAPI
  • Next.js
  • Supabase
  • Inngest
  • Sentry
  • Vercel