How to Check If AI Chatbots Know About Your Business (Free Tool + Manual Method)

Published 1 September 2026 · Updated 3 September 2026 · 7 min read

You rank on page one of Google. ChatGPT has never mentioned you. When someone asks an AI assistant a question your business should answer, a competitor’s name comes back instead. This is not bad luck — it is a technical gap, and it is diagnosable.

This guide walks through both methods for checking your AI visibility: the manual approach (querying the AI systems directly) and the technical approach (checking the signals that determine whether AI can find and cite you in the first place).

Why Google Rank Does Not Predict AI Visibility

Before running any test, it helps to understand why SEO rankings and AI visibility measure different things. Search engines rank pages based on authority, relevance, and user signals. AI systems decide what to cite based on:

• Whether their crawlers can access your site (robots.txt allows ChatGPT-User, OAI-SearchBot, Claude-User, PerplexityBot, etc.)
• Whether your content is readable as text (not locked in JavaScript bundles)
• Whether you have structured data that declares what your pages are about
• Whether your content directly and accurately answers the question being asked
• Whether you have clear entity signals (consistent brand name across pages and schema)

A site with high domain authority but JavaScript-rendered content and no schema markup may score 30/100 for AI visibility while a newer site with clean HTML and proper structured data scores 90/100. The technical signals matter independently of backlink authority.

Method 1: The Manual Query Test

The quickest way to get a directional answer is to ask the AI systems directly. Use these query types across ChatGPT, Claude, and Perplexity:

Brand queries

Ask directly: “What is [your brand name]?” and “Tell me about [your brand name] — what do they do?” If the AI has accurate information, it knows about you. If it says it cannot find information or gives an incorrect description, your entity signals are weak.

Topic queries

Search for your primary topic: “What are the best [your category] tools?” or “How do I [the main problem you solve]?” If competitors appear and you do not, you are losing AI-referred queries in your category. Note which competitors are cited — examine their technical setup for what you are missing.

Comparison queries

Ask: “Compare [your brand name] vs [a competitor].” If the AI knows the competitor well but has no data on you, that is a strong signal that your site is either not crawlable by that AI system’s bots, or not producing content the AI can confidently cite.

What to note

Record which AI systems mention you, what they say (accurate vs. inaccurate vs. nothing), and which competitors they cite in your category. This gives you a directional baseline. It is not a statistically valid measurement — single-query results from live AI systems vary by run — but it identifies obvious gaps.

Method 2: The Technical Check (18 Signals)

A manual query test tells you the symptom. The technical check tells you the cause. There are 18 deterministic signals that determine whether AI systems can find, read, and cite your site. They fall into five categories:

1. AI Crawler Access

The first thing to check is whether your robots.txt allows the major AI crawlers. If you are blocking the live-answer crawlers — ChatGPT-User and OAI-SearchBot (OpenAI), Claude-User and Claude-SearchBot (Anthropic), PerplexityBot and Perplexity-User — those systems cannot fetch your content to answer a question, and Aura scores it as a fail. Blocking only training crawlers such as GPTBot, ClaudeBot or Google-Extended is scored as a warning. Check your yourdomain.com/robots.txt directly and look for:

• User-agent: ChatGPT-User and User-agent: OAI-SearchBot — should be followed by Allow: /, not Disallow: /
• User-agent: Claude-User and User-agent: Claude-SearchBot — same check
• User-agent: GPTBot — same check (training crawler; a block here is a warning)
• User-agent: ClaudeBot — same check
• User-agent: PerplexityBot — same check
• User-agent: Google-Extended — same check

A blanket User-agent: * with Disallow: / blocks everything, including AI crawlers. This is the single most common cause of zero AI visibility in otherwise well-built sites.

2. Machine Discovery

Three files signpost your content to AI systems. The scan checks that each is present — it does not evaluate their contents:

• yourdomain.com/llms.txt exists — the single largest Machine Discovery check (8 of 15 points); see our llms.txt writing guide
• yourdomain.com/sitemap.xml returns an XML sitemap (4 points; presence only — whether robots.txt references it or which pages it lists is not scored)
• Your pages expose an RSS/Atom feed link (3 points)

3. Structured Data

Schema.org markup tells AI systems what type of content each page contains. Check your key pages using Google’s Rich Results Test or a JSON-LD validator. The schema types the scan awards points for:

• Organization, WebSite or LocalBusiness on your homepage — declares your entity clearly
• Article or BlogPosting on blog content — signals citable editorial content
• FAQPage on FAQ or support pages — feeds AI answer extraction
• Valid JSON-LD syntax, Open Graph title and description, and a canonical tag round out the category. Product, Service and HowTo markup can be useful but earn no points in the scan.

4. Content Citability

AI systems prefer content that is structured for comprehension, not just optimized for visual design. Aura reads the raw HTML and does not execute JavaScript, so anything that only appears after JS runs is invisible to the scan. Check:

• The page title is 15–70 characters and the meta description is 50–170 characters
• Exactly one H1 and at least two H2s — AI systems use headings to navigate content
• At least two lists or tables — the blocks AI answers most readily extract
• At least 300 words of visible text (120–299 earns partial credit)

5. Technical Trust

These baseline signals determine whether a crawler can fetch and index the page at all:

• HTTPS — the page is served over HTTPS (the scheme is checked; the certificate is not validated)
• Response time — under 1.5 seconds for full credit, under 4 seconds for partial credit
• HTTP 200 — the final URL after any redirects returns a 200 status

Method 3: The Automated Scan (Fastest)

Checking all 18 signals by hand means reading robots.txt, the sitemap and llms.txt, then counting headings, words and schema blocks in view-source yourself. The Aura free AI-visibility scan runs the same 18 checks in about ten seconds. Enter your domain and you get:

• A score from 0 to 100 across all 18 checks, grouped by category
• A pass, warn or fail verdict on each individual check — no guessing which ones you failed
• A specific fix for every failing check — not generic advice, but the exact action to take for your site
• A letter grade (A through F) that gives you a single number to track over time

No account required. No email gate. Run it for your domain, then run it for a competitor to see the gap.

What to Do With Your Results

If your score is below 65 (a D or an F), start with the two checks that have the highest impact and the lowest effort to fix:

1. AI crawler access — if any major AI bot is blocked in robots.txt, fix it first. One line change, immediate impact on crawlability.
2. Structured data — add an Organization schema to your homepage if you do not have one. Paste the JSON-LD block into your <head> and validate it. One hour of work, lasting benefit.

If your score is 65–89 (a C or a B), focus on Content Citability and the remaining Structured Data checks — heading hierarchy, lists and tables, canonical URLs, Article or FAQPage schema on your key pages.

If your score is 90 or above (an A) and AI still is not citing you in your category, the gap is usually content depth, not technical signals. AI systems cite sources they can confidently attribute a fact to. Producing more specific, factual, directly-answerable content matters more at that level than further technical optimization.

Tracking Over Time

AI visibility is not a one-time fix. New content pages can accidentally block crawlers. Schema markup can break during site rebuilds. AI systems update their crawlers and citation behavior over time.

The Monitor tier of Aura runs your 18-check scan on a weekly schedule and emails you when your score moves by 8 or more points, your grade changes, or a live-answer crawler is newly blocked (smaller moves are logged on the dashboard) — so you know quickly if a robots.txt change broke your AI crawlability, rather than discovering it weeks later when queries stop returning your brand.

Frequently Asked Questions

How do I know if ChatGPT knows about my website?
Ask ChatGPT directly about your brand and topic. If it cannot find information or describes you inaccurately, run a free AI-visibility scan to find the technical cause.

Why does AI cite my competitors but not me, even though I rank higher on Google?
AI citation depends on technical signals — crawlability, structured data, content in plain text — not on domain authority or backlink count. See the full breakdown.

Can I make AI always cite my website?
No. Citation is determined at query time by the AI. You can maximize your chances by scoring well on all 18 technical signals and producing content that directly answers the questions your audience asks.

What is the fastest way to check my AI visibility?
Run the free scan at aaivisibility.com. Enter your domain, get a score across 18 checks in about ten seconds. No account required.

Check your AI visibility score — free, takes 30 seconds