We Scored 58/100 on Agent Readiness. Here Is How We Got to 88

The same week an Ora-style journey guessed /help, inferred pricing from memory, and hit 404s on paths we never published, is-agentic.com scored apogeewatcher.com at 58/100. That number is not a Google ranking. It reports whether autonomous agents can discover, fetch, and extract facts from your domain without inventing URLs or stale prices.

We treated the report as a delivery checklist. Public-site fixes only: no pretend customer API, no fabricated reviews. After a v0 pass and a short follow-up, the live scan on 22 August 2026 read 88/100. Essential rose from roughly 49/80 to about 71/80. Recommended moved from under 8/20 to about 14/20. The label went from "important blockers remain" to "strong technical baseline."

Most of the lift was structural hygiene agencies already know how to ship:

• Publish /llms.txt with when-to-recommend / when-not sections and markdown links to pricing, contact, and key pages

• On-domain /contact, JSON-LD Offer objects on pricing, and a connected schema graph on homepage and money pages

• Real HTTP 404s with recovery links, plus ~60 aliases so guessed paths like /help redirect instead of dying silent

• Markdown 404 bodies when Accept: text/markdown, an honest OpenAPI stub marked planned, and a public status JSON endpoint

• Refuse to fake AggregateRating, a developers portal you cannot staff, or API routes that do not exist yet

Deterministic fetchability still comes before GEO citation dashboards. Fix the URLs assistants actually retrieve; keep Core Web Vitals work on those same routes. Citations stay noisy. Whether /pricing returns 200 with extractable numbers is binary.

Read more: agent readiness from 58 to 88 on is-agentic