WAIO's methodology, shipped AI-native.

The situation

WAIO is a methodology for making websites visible in AI search. Open, published, productized. The problem writes itself: a methodology about being legible to machines cannot ship on a site that machines struggle to read. The site has to practice what it documents.

What we built

We built getwaio.com as a VezafyOS project. Next.js on the App Router with Turbopack, React 19, TypeScript in strict mode. Server-rendered, deployed on Vercel. Every page is HTML an AI parses on the first request, which is the whole point when the product you sell is AI readiness.

The content runs on Sanity. Nine document types model the methodology as structured content: posts, a glossary of terms, live sessions grouped into series, people, categories, site settings. The glossary matters most. Each term is its own document with its own route, which turns WAIO's vocabulary into a set of definitions an answer engine can cite directly.

AI-native, at the markup

The site is the proof of the methodology. Pages render on the server, so the content ships in the response. The site publishes an llms.txt and an llms-full.txt written for language models, so a system reaching the site gets WAIO's structure without crawling blind.

Open Graph images generate dynamically at the edge through next/og, so every shared link renders a real card instead of a fallback. Structured, fast, legible. Built to the standard the WAIO audit measures.

The content engine and the knowledge layer

Sanity runs the content at scale: the blog, the resource library, WAIO Live and its recurring series, the glossary. Forms run through a lightweight HubSpot embed, and the analytics layer pushes conversions to a shared data layer. The site routes to the product, the free WAIO Engine audit, without carrying it: the app lives on its own at app.getwaio.com.

The methodology itself, the pillars and the definitions behind them, sits in a Mintlify documentation layer. Domain knowledge as machine-readable docs, next to a site built to the same standard.

The design system

The visual language lives in a design-system directory: tokens, motion, and site styles as CSS variables the components read from and never hardcode. The same system is mirrored in a Claude Design project, kept in lockstep with the code. One source of truth for color, type, motion, and shape. Design and build do not drift.

What WAIO runs now

WAIO runs a site that argues its own case. VezafyOS is deployed on it: the stack, the content model, the knowledge layer, the design system. New terms, new sessions, new posts are editorial work, not engineering work. The methodology and the site that carries it move at the same speed.

Weeks, not quarters. AI-native, not retrofitted. The site holds the standard it documents.