From the founder
The Blueprint exists because the gap between agentic capability and product trust is real, growing, and currently filled by static documentation, vendor-locked patterns, and people figuring it out alone. This page is a short note on why; the work itself is the most useful artefact.
Why this exists
After seventeen years working on product trust questions in different forms, most recently leading AI strategy across design, research, and brand at a travel company in the UK, I started building this as the version of that work that didn't fit a day job.
There's a quieter reason alongside the strategic one. I wanted to take a 0-to-1 attempt through alone, to find out what one person's reading of an emerging discipline produces when no team shapes the work. The platform might become a business, or it might end as evidence of what the bet was worth. Both outcomes are honest. The work has value either way.
How the work gets made
The thing I'd most want to demonstrate about the Blueprint isn't the principles. It's the loop the platform applies to itself.
A brand agent audits every pull request against a numbered catalogue of voice smells before merge. The validator that scores other people's agentic architectures runs against the platform's own source code, with findings folded back into the next iteration. Every readiness score on the site links to a public run-id with the full iteration history visible to anyone who clicks through.
The discipline the doctrine asks of others, the platform applies to itself first. The loop is visible in the codebase, the audit reports, and the public review pages.
What's on the bench
A page surfacing the audit loop publicly, so the recursive integrity is legible without reading the source. More editorial in field-report register. Continued presence in builder communities where the wedge is being tested in real conversations.
The architecture is broader than the surface area currently visible. Page-by-page coverage will catch up over time. This is built in evenings and weekends, by one person, alongside a full-time role. The pace reflects that, and so does the rate of change. Both are part of the bet.
How to reach out
Practitioner conversations are welcome on LinkedIn or GitHub. The closed beta is the route in for builders who want validator access. The most useful threads are one of two kinds: "we shipped something the doctrine should know about," or "this part doesn't hold up, here's why." Both compound the work.
Run by agents. Governed by humans. The page you're reading was audited by the same brand agent that reviews every other surface on the site.