METHODOLOGY

How We Audit and Monitor AI Visibility

Full transparency on our process, tools, scoring method, limitations, and what you can expect from our services.

Platforms Tested

We test on ChatGPT (OpenAI), Google Gemini, Perplexity, Claude (Anthropic), and Microsoft Copilot. These are the most widely used AI assistants as of mid-2025.

Questions Analyzed

For each audit, we identify 15-20 real buyer questions — the kind potential customers actually ask AI engines about your industry, products, or services. Each question is tested on 3 platforms, producing 45-60 analyzed responses per audit.

Languages and Countries

We test in the language your customers use. Most audits are conducted in English, but we also support French and can adapt to other languages on request. Location-specific queries are used when relevant (e.g., 'best real estate agency in Kuala Lumpur').

Frequency of Tests

The Full Audit is a one-time analysis. The Monitoring plan repeats the analysis monthly, tracking changes in how AI engines cite your business over time.

How We Assess Visibility

For each AI response, we check: Is your business mentioned? Is the information accurate? Is the sentiment positive, neutral, or negative? Are you recommended or just listed? We compare results with 3 direct competitors. We do not use a proprietary score algorithm — we report observable facts.

Technical Audit

We check: robots.txt AI crawler access rules, llms.txt presence and format, structured data (JSON-LD / Schema.org), sitemap accessibility, server response to AI bot user agents, CDN or firewall bot blocking.

Variation of AI Responses

AI responses are not static. The same question asked on the same platform can produce different answers at different times. Our observations are snapshots, not permanent states. We report what we observe on specific dates.

No Guarantee of Placement

No company can guarantee that an AI engine will cite, recommend, or mention a specific business. We improve the technical signals and structured information that AI engines use. We do not control their outputs.

Correction Policy

If a delivered report or implementation contains errors, we correct them at no additional cost. Our goal is accuracy and usefulness.

Delivery Times

Free Check: within 24 hours. Full Audit: within 72 hours. Done-for-You Setup: 5-7 business days. Monitoring: monthly report delivered within the first week of each month.

Get Your Free Check