AI visibility audit for ChatGPT and AI search
I measure which prompts get your brand named by AI systems, who they recommend instead and which sources they rely on. You get a log of every tested prompt, a list of issues and a prioritized plan for the next 30–90 days.
Two green out of five is a normal start. The problem is not knowing which two.
When does an audit make sense?
If any two of these sound familiar, measure where you stand before you pay for fixes.
AI names your competitors
You ask ChatGPT about your category and get a list of competitors and directories. Your brand is not on it.
The facts are wrong
A model quotes an old price, a closed location or a service you stopped offering years ago.
Priorities are unclear
There are a dozen ideas about what could be done, but no data on which of them will change the result.
You need a comparison
You want to see in numbers how your visibility differs from your competitors’. The extended audit benchmarks 5 of them.
AI traffic is invisible
Visits from AI systems are not separated in your analytics, so there is no way to judge the channel. The audit starts one step earlier and shows whether the answers name you at all.
You are about to start work
You need a baseline before the fixes begin, so that three months from now there is something to compare against.
What goes into
the audit?
Prompt set
- Buying prompts: “who can…”, “best company for…”, “how much does it cost”
- Informational questions from your field
- Brand prompts: what a model says when asked about you directly
- Comparison prompts: “X vs Y”, “top providers of…”
- Built from your Google Search Console queries, sales questions and SERP analysis
- You review and approve the final list before testing starts
Testing across AI systems
- Standard: 4 systems chosen with you – from ChatGPT, Google AI Overviews and AI Mode, Gemini, Perplexity, Microsoft Copilot, Claude and DeepSeek
- Extended: all 7
- Each prompt runs several times, because answers change between runs
- Brand mentions, domain citations and position in lists, recorded for every answer
- The tone and context in which your brand comes up
- Who the systems name instead of you, and how often
- Screenshots and the full text of every answer
Competitor benchmark (extended audit)
- Share of voice for 5 competitors in each system
- Which sources the answers about them are built from
- What makes those sources easy for a model to cite
Gap analysis: citations, content, entity
- Citation gap: sites where competitors are cited and you are absent
- Content gap: topics and questions your site does not answer
- Entity gap: missing or conflicting facts about your company across the web
Quick technical check
- Indexing and access to key sections, checked in Google Search Console and Bing Webmaster Tools
- Whether the main content is in the HTML without running JavaScript
- AI crawler access: robots.txt, response headers, firewall and CDN blocks
- Markup, heading structure and blocks a model can extract as an answer
Fact accuracy
- Prices, addresses, services and contacts in AI answers compared with your real data
- The sources that feed outdated information to the models
Priorities
- Every task scored by impact, effort and urgency
- Split into tasks I can take on and tasks for your team
- A roadmap for the next 30–90 days
How does the project run?
Brief and access
You tell me about your market, regions and competitors and share read access to analytics. 1 day.
Prompt list
I send the prompt list, you add to it and approve it. 1 day.
Testing
I run the prompts in the chosen systems and record every answer and source. 2–3 days.
Analysis
Gap analysis, technical check and task scoring. 1–2 days.
Report
You get the files and the roadmap, and they are yours to keep.
Walkthrough call
40–60 minutes on what to do first and what each step should change.
What do you get?
All files stay with you, and any contractor can work from them.
Prompt and answer log
Every prompt, each system’s answer, brand mentions, domain citations and sources.
List of cited sources
The sites and competitors the models rely on in your market.
Issue list
Technical, content and off-site obstacles, each with a note on what it affects.
30–90 day roadmap
Tasks in priority order, with effort estimates and an owner for each.
Summary deck
A short version for the business owner or leadership team.
Walkthrough call
Findings and your questions, 40–60 minutes, recorded.
You can hand the log to any contractor. It describes your market, whoever ends up doing the fixes.
Timeline and format
- Standard audit – 5 business days from when I get access
- Extended audit – up to 7 business days
- Questions along the way by email, Telegram or WhatsApp
- The walkthrough call within a week of delivery, at a time that suits your time zone
What I need from you
- Your services, markets and main competitors
- Read access to Google Analytics 4 or the analytics tool you use
- Google Search Console and Bing Webmaster Tools, if they are set up
- Current prices, addresses and contacts to check the facts against
Not included
- Fixing what the audit finds – that is technical GEO, page optimization or a monthly plan
- Writing new content
- Repeat measurements later – available as monitoring and included in the monthly plans
How is the result measured?
The audit sets the baseline. After the fixes, the same numbers are measured again on the same prompt set.
Prompt coverage
Share of prompts in the set where your brand appears in the answer.
Domain citations
How often your site is listed among the sources of an answer.
Fact accuracy
Share of answers that state your company’s facts correctly.
A Telegram store builder named first of 8 in ChatGPT
At the start, none of the 4 tracked AI systems named the service. After technical fixes, extractable facts and an llms.txt file, all 4 name it; it took 90 days to appear in answers. Russian-speaking market.
What usually comes after the audit?
Technical GEO
If the audit shows AI crawlers cannot reach your content, this is where to start.
from $490 →Page optimization for AI search
Existing pages reworked into answers a model can extract and quote.
from $90 per page →AI visibility monitoring
Regular checks on the same prompt set, with a monthly report I write myself.
from $290/mo →Related reading
How to check your brand's visibility in ChatGPT, Gemini and Copilot
One prompt with your brand name proves very little. Here is the repeatable method I use: a fixed prompt set, repeat runs, a log of full answers, AI share of voice and what Bing and Google report.
Read →What is generative engine optimization (GEO), and how is it different from SEO and AEO?
GEO is the work of getting your business named and cited in AI answers. I explain how it relates to SEO and AEO, what to measure instead of rankings and what a realistic first 90 days look like.
Read →Questions about the audit
How is this different from an SEO audit?
Which AI systems do you check?
Are the results reproducible?
Can I order just the audit?
How many prompts do I need?
Do you guarantee AI will name me after the fixes?
Order an AI visibility audit
Send your website and a line about your market. Within one business day I will reply with a draft prompt list and an exact timeline.
I reply within one business day. For a larger project, fill in the project brief.