"AI visibility score" is not a standard metric. Different tools use the name for two different things, and it's worth knowing which one you're looking at before you act on the number:
- Outcome scores measure how often AI assistants mention or cite your brand when people ask questions. They come from sending many prompts to ChatGPT, Gemini, Google AI Mode and others and counting how often you appear, usually compared with competitors.
- Readiness scores measure whether AI crawlers and agents can reach, read and understand your site at all. They come from auditing your pages: robots.txt, rendering, structured data, llms.txt and similar.
Outcome scores tell you where you stand. Readiness scores tell you what to fix. A site that blocks AI search bots or renders nothing without JavaScript will struggle to get cited, whatever its content says, so the two are related, but a high readiness score doesn't guarantee mentions.
How RankRoot's AI visibility score works
RankRoot's score is a readiness score. The free audit fetches your homepage plus /robots.txt, /sitemap.xml, /llms.txt, /llms-full.txt, /openapi.json, /.well-known/mcp.json and /.well-known/ai-plugin.json, then runs 22 checks in four categories. Each check has a weight. A pass earns the full weight, a warning earns 40% of it and a fail earns nothing. Your score is the points earned divided by the 110 points available, as a percentage. Each category score is calculated the same way using only that category's checks.
| Category | Checks | Points | Share of score |
|---|---|---|---|
| AI crawler access | 7 | 38 | 34.5% |
| Answer-engine content | 8 | 32 | 29.1% |
| Agent readiness | 4 | 25 | 22.7% |
| Structured data | 3 | 15 | 13.6% |
| Total | 22 | 110 | 100% |
All 22 checks and their weights
| Check | Category | Weight |
|---|---|---|
| AI search & answer bots allowed (OAI-SearchBot, PerplexityBot, Claude-SearchBot, Claude-User, Googlebot, Bingbot) | AI crawler access | 16 |
| Content visible without JavaScript (300+ words in raw HTML) | Answer-engine content | 12 |
| llms.txt | Agent readiness | 9 |
| JSON-LD structured data | Structured data | 7 |
| Machine-usable interface (OpenAPI, MCP, ai-plugin.json or API docs link) | Agent readiness | 7 |
| Page is indexable (no noindex) | AI crawler access | 5 |
| Entity / product schema (Organization, Product, SoftwareApplication...) | Structured data | 5 |
| Pricing is machine-readable | Agent readiness | 5 |
| AI training crawlers allowed (GPTBot, ClaudeBot, Google-Extended...) | AI crawler access | 4 |
| XML sitemap | AI crawler access | 4 |
| Descriptive title tag (15-70 characters) | Answer-engine content | 4 |
| Meta description (50-170 characters) | Answer-engine content | 4 |
| Docs / help content linked | Agent readiness | 4 |
| Served over HTTPS | AI crawler access | 3 |
| robots.txt present | AI crawler access | 3 |
| Fast response (under 1.5 s) | AI crawler access | 3 |
| One clear H1 | Answer-engine content | 3 |
| Scannable section headings (3+ H2s) | Answer-engine content | 3 |
| Question & answer content | Answer-engine content | 3 |
| Linked brand profiles (sameAs or 2+ social links) | Structured data | 3 |
| Open Graph tags | Answer-engine content | 2 |
| Language declared | Answer-engine content | 1 |
The two heaviest checks are the ones that can make you invisible on their own: AI search bots blocked in robots.txt (16 points) and a page that's empty without JavaScript (12 points). ChatGPT-User is reported but doesn't cost points, because OpenAI says it acts on user requests and robots.txt rules may not apply to it. Blocking training crawlers is only a warning, since it's a legitimate choice (see should you block GPTBot?).
Grades
| Grade | Score |
|---|---|
| A | 90-100 |
| B | 75-89 |
| C | 60-74 |
| D | 40-59 |
| F | Below 40 |
You can see real scores for 500+ SaaS and AI products in the AI Visibility Index, each with its own page listing the issues found. Response time is measured from our crawler and is noisy in batch runs, so it's left out of the issue lists on those pages.
How it differs from Semrush's AI visibility metrics
Semrush uses the same name for an outcome metric. Its knowledge base defines the AI Visibility Score in the AI Visibility Toolkit as "a benchmark score (0–100) showing how often your brand appears in AI-generated answers compared to competitors", benchmarked against the median number of mentions for your top industry competitors, which Semrush identifies automatically. Semrush's separate Site Audit has an AI Search Health score, which is closer to a readiness score.
| RankRoot AI visibility score | Semrush AI Visibility Score | Semrush AI Search Health | |
|---|---|---|---|
| Measures | Whether AI crawlers and agents can reach, read and understand your site | How often your brand appears in AI answers compared with competitors | How optimized your site is for AI search |
| Based on | 22 weighted checks of your homepage and root files | Mentions across Semrush's prompt database | AI search checks, AI bot access in robots.txt and technical readiness (schema, crawlable navigation, internal links, llms.txt) |
| Method published | Yes: every check and weight is on this page | Definition published; the formula isn't | Inputs listed; scale and weighting aren't documented |
| AI platforms | Not platform-specific: checks access for ChatGPT, Perplexity, Claude, Google and Bing crawlers | Prompt database: ChatGPT, Gemini, Google AI Overviews and AI Mode | Checks against eight AI crawlers |
| Output | Score, grade, issues with fixes, and generated llms.txt, robots.txt and JSON-LD | Score plus mentions, citations, share of voice and sentiment reports | Score plus AI search issues in the Site Audit Issues tab |
| Price | Free scan; Traffic Boost $49 once (30-day featured launch plus 7 days of Pro for up to 50 pages per scan); done-for-you AI Visibility Employee from $149/mo | AI Visibility Toolkit: $99/mo, no free trial | In Semrush Site Audit; the toolkit includes AI search checks for up to 100 pages |
Semrush says its prompt data is captured from real requests rather than LLM APIs, comes from over 317 million prompts and responses, and is updated daily on a rolling basis. Its Brand Performance reports (share of voice and sentiment) cover Google AI Mode, ChatGPT, Perplexity and Gemini and update weekly. If you need that kind of tracking across many prompts, Semrush is built for it. If you want to know why you're missing and get the fix files, a readiness audit is the faster first step. See the full Semrush AI Visibility Toolkit alternative comparison.
How to check your AI visibility
- Readiness: run the free AI visibility audit. It takes about ten seconds, needs no signup and shows your score, grade, category scores and every failed check with a fix.
- Mentions: run the AI mention checker. It asks Google Gemini the questions your buyers ask and shows whether you're mentioned, where you rank, which competitors it names instead and which sources it cites. The free check runs 3 prompts a day.
- By hand: ask ChatGPT, Perplexity and Claude your buyers' questions and log the answers. Our guide on checking if ChatGPT mentions your brand has a tracking template.
How to raise your score
Fix the heaviest failures first. Allow AI search bots in robots.txt and in any CDN "block AI bots" setting (robots.txt for AI crawlers), make sure the page's text is in the server-rendered HTML, publish an llms.txt and add Organization and SoftwareApplication JSON-LD. The audit generates the llms.txt, robots.txt and JSON-LD for you. Then rescan to confirm.
Sources
- Semrush KB: AI Visibility Toolkit (read 2026-10-17)
- Semrush KB: AI Visibility metrics (read 2026-10-17)
- Semrush KB: Where does the data in the AI Visibility toolkit come from? (read 2026-10-17)
- Semrush KB: AI Search Health in Site Audit (read 2026-10-17)
- OpenAI: Overview of OpenAI crawlers (read 2026-10-17)