Is it bad if our own site scores low?

For a launch-stage product, a low score with passing technical checks is common. AI recommends brands with entity clarity, third-party mentions, and buyer-intent content — not just a clean robots.txt file.

What passed on our audit?

All AI crawlers allowed, homepage indexable, JSON-LD schema present, FAQ content detected, and llms.txt available. The gap was authority and entity recognition, not basic technical blocking.

Category and recommendation prompts surfaced established competitors instead of us. Branded queries sometimes confused our product name with generic phrases or unrelated brands.

What are we doing about it?

We published definitive product and methodology pages, a buyer's guide, and this case study. Next up: third-party mentions and comparison content in formats AI already cites for our category.

← All posts

19 August 2026

We audited ourselves at 8/100 — what a new AI visibility product should expect

AI Can't Find Me scored invisible on its own diagnostic. Technical checks passed; category recommendations didn't. Here's what we found and what we're fixing.

We sell a one-time AI visibility diagnostic. So we ran it on ourselves.

Score: 8/100 (invisible). That sounds embarrassing until you look at the evidence — and at how young the product is.

What passed

Our technical baseline was solid:

GPTBot, OAI-SearchBot, PerplexityBot, ClaudeBot, and Bingbot — all allowed
Homepage indexable, no snippet blocks
JSON-LD schema detected
FAQ-style content present
llms.txt at the site root

If you ran only our Free AI Crawler Check or AI Readiness Score, you'd think we were in good shape.

What failed

The live prompt panel told a different story.

On category queries — "best AI visibility audit in Sydney", "who should I hire" — we were 0 for 6. Perplexity named LoudScale, Roxane Pinault, and other established players. ChatGPT surfaced AISearch Global and similar agencies.

On branded queries, results were inconsistent. Sometimes ChatGPT described us as a service; sometimes Perplexity said it couldn't verify reviews; occasionally the name was treated as a generic phrase rather than a product by kduai.

That's the gap our report is built to find: crawlable ≠ recommendable.

Why this is normal for a new product

AI systems recommend what they can confidently identify as a real entity with third-party proof. A two-week-old domain with strong technical setup but no directory listings, comparison pages, or external mentions is invisible on buyer-intent prompts — even when the product works.

We didn't build this report to flatter ourselves. We built it to show the difference between a crawl score and what a customer actually hears when they ask ChatGPT who to hire.

What we shipped after the audit

What is AI Can't Find Me? — definitive product entity page
Report methodology — citeable explanation of how testing works
How to choose an AI visibility audit — buyer's guide for category queries
This post — public dogfooding

What we're doing next

Earn third-party mentions on pages AI already cites for our category
Publish comparison content in listicle formats engines prefer
Re-run the diagnostic in 30 days and publish the delta

If you're in the same boat — technical checks green, AI still naming competitors — that's exactly what the full diagnostic is for. We'd rather show you an honest 8/100 with evidence than a vanity score with no fix plan.