Writing
Field notes on AI back-office engineering, not "AI news."
Most AI writing is summary of someone else's work. We don't do that here. Every piece on this page comes from first-hand engineering — something we built, broke, or measured. Mostly broke.
First essay drops mid-June 2026. Three pillars in the pipeline:
- Episode 001 P-010 In production
Content starvation is the AI consulting market's default failure mode
Why we built our own content pipeline before we ever wrote a sales page. And what broke on day two.
- Episode 002 P-001 In production
Why our NDIS claim system was designed never to see a participant's name
The PII-isolation pattern, the audit it survived, and the engineering compromise that made it shippable.
- Episode 003 P-004 In production
The first system we ship always breaks on day two
A field guide to the failure surface in allied-health clinical-notes systems. From a real failed v1.