Why is my AI Readiness score low?
Low AI Readiness tool scores reflect missing evidence across five categories: bot discoverability (30 points), structured data (25), LLM communication (25), content accessibility (15), and cross-signal readiness (5). Review the category breakdown for lost points. This tool score differs from the overall Radar dashboard average.
Radar has two distinct scores. The dashboard overall score is the unweighted average of completed audits with usable scores, excluding withheld grades. The individual AI Readiness Score is one audit inside that run. It sums five categories with unequal maximum points: 30 + 25 + 25 + 15 + 5 = 100. The five categories are not equally weighted.
The 5 categories scored
| Category | Max points | What it checks |
|---|---|---|
| Bot discoverability | 30 | Access for supported AI browse and search bots, robots.txt and sitemap declarations, deliberate AI training policy, and meta robots restrictions |
| Structured data | 25 | JSON-LD presence, entity and navigation schema, richer schema types, and schema diversity |
| LLM communication | 25 | llms.txt presence, structure, entity definitions, links, use policy, content depth, and llms-full.txt |
| Content accessibility | 15 | Server-rendered content, title and description length, content volume, and SSR-first framework signals |
| Cross-signal readiness | 5 | Structured data paired with llms.txt, AI bot access paired with rendered content, and absence of high-priority recommendations |
Highest-impact fixes by failing category
- Bot discoverability low — review the supported browsing and search bots in your findings, resolve unintended access restrictions, and add a Sitemap directive. Keep an explicit policy for training bots.
- Structured data low — add Organization JSON-LD to root layout. Add Article schema to blog pages. Validate via Rich Results Test.
- LLM communication low — ship a complete /llms.txt with brand identity, product list, key URLs, FAQ, and use policy.
- Content accessibility low — ensure server-side rendering, use semantic HTML (h1-h6, nav, article), avoid JS-only content rendering.
- Cross-signal readiness low — pair structured data with llms.txt and AI bot access with server-rendered content, then resolve the high-priority recommendations shown in the audit.
When to check the score again
Re-run the audit after deploying your fixes and check the evidence collected for that run. Score changes depend on which checks changed and whether the audit could access the updated content. An audit score does not establish when an AI provider will re-crawl or cite your site, and there is no fixed score increase or provider timeline promised here.
Prioritize the missing checks shown in your category breakdown. Their maximum points differ, and cross-signal points depend on combinations of evidence. If a scan cannot collect enough evidence, its score may be withheld rather than graded.
How to re-baseline after fixes
Re-run Radar AI Readiness after each batch of fixes. The category breakdown shows which fixes moved the needle. Compare to the prior audit for delta tracking.
Want to take action on this?