Service
Agent Accessibility Audit
Can AI assistants find, understand and recommend your business? We score it out of 100, then show you what to fix.
Why it matters
More customers now ask an AI assistant who to hire, where to go or what to buy. The assistant answers from what it can read and verify on your website. When something is missing, it guesses, or it recommends someone it can verify.
Three questions
- Can agents find you? Crawl access, sitemaps, canonical URLs, share images, llms.txt.
- Can agents understand you? What you offer, for whom, where, who's behind it, and proof.
- Can agents do something for your customers? Clear next steps, working forms and links, contact details.
What we score
| Dimension | Weight |
|---|---|
| Discoverability | 10 |
| Machine readability | 15 |
| Knowledge clarity | 15 |
| Context & intent | 10 |
| Trust & authority | 10 |
| Conversational readiness | 10 |
| Actionability | 15 |
| Brand representation | 5 |
| Safety, boundaries & failure | 5 |
| Freshness & maintainability | 5 |
What you get
- A 100-point score with a plain-language verdict
- An agent test: real customer questions, and what an AI can (and can't) truthfully answer from your site today
- Findings for each dimension, with effort and impact for every fix
- Prioritized fixes: critical, high value and quick wins
- A 30-day plan, plus a one-page summary to share with your team
Case study: Stingray Scout Drone Services
47 → 78
Stingray Scout Drone Services gained 31 points in one round of fixes.
The biggest gains came from basics most sites skip: a robots.txt file, a sitemap, canonical tags, a share image, an llms.txt summary for AI tools, and a privacy policy for the contact form.
Or email hello@radishimmersive.com.