Townsend AI Studio
A family health startup · Assessment · September 2026HealthThe free assessment

The best thing they'd built wasn't shipping

346
verified records
0
users reached
3
phases scoped

The situation

A founder-led startup with a public website and, underneath it, a content pipeline that was genuinely well engineered: 346 records, each independently verified with a full source trail. The site was live. The pipeline was done. They wanted to know what to fix first.

What I found

The best thing they'd built wasn't shipping. The 346 verified records reached zero users — the public site still served the earlier, unverified content, because the last stage of the pipeline was never built. Around it: 1,300+ tests that ran nowhere, about 15 recent merges that bypassed review, a public form that needed hardening, and a rules file for their AI tooling that had been silently stripped for six weeks. Their rules were good; nothing enforced them.

What came out of it

A written readout. A punch list they could do themselves, for free: wire the tests into CI, pin the dependencies, harden the public form, resolve the auth setup. And a scoped three-phase plan for what needs specialist hands — guardrails, a rebuild where the verified corpus is the only content layer, and a pipeline the founder runs solo.

The number

346 verified records, reaching 0 users.

← All work

Availability

Have something like this?

The assessment is where this case started. Two calls, free, one per company.

Start with a free assessment →
Reply: 2 business daysSeattle & remote