Does AI recommend your business? A 30-query audit you can run yourself
To find out whether AI assistants recommend your business, run the same 30 customer-style questions across ChatGPT, Gemini, Perplexity, Microsoft Copilot, and Google AI Overviews, record which businesses each one names, and note the sources it cites. Most local and regional businesses appear zero times on their first audit. That is not a failure. It is a baseline, and the sources the AI cites instead of you are your to-do list.
We ran this exact audit on our own agency earlier this year. Thirty queries, five platforms, zero mentions of ACS Creative. Here is the method, so you can run it on your business in about two hours.
Why this matters now
AI assistants now answer questions customers used to type into Google. “Who is a good web design agency near Frederick?” or “best pediatric dentist in Rockville” returns a short list of named businesses, with citations, and no page two. If you are not on that list, you do not exist for that customer. Traditional rankings do not transfer automatically. A business ranking third on Google can be absent from every AI answer, and vice versa.
Step 1: Write your 30 queries
Split them into three groups of ten.
Discovery queries. How a customer who does not know you would ask. “Best [service] in [city].” “[Service] company near [neighborhood].” “Who should I hire for [problem] in [county]?”
Comparison queries. “Top [service] providers in [region].” “[Competitor] alternatives.” “[Service] agencies that specialize in [industry].”
Brand queries. “What is [your business]?” “Is [your business] good?” “[Your business] reviews.” These test whether the AI knows you at all and what it says when asked directly.
Use the language customers use, not the language on your website. If you sell “digital experience solutions,” your customers ask for “a website.”
Step 2: Run each query on five platforms
- ChatGPT (with search enabled)
- Google Gemini
- Perplexity
- Microsoft Copilot
- Google AI Overviews (a normal Google search; note whether an AI Overview appears and who it names)
Use a logged-out or fresh session where possible so your own history does not influence results. Run each query once per platform. That is 150 answers.
Step 3: Record four things per answer
Build a spreadsheet with one row per query and platform:
- Were you named? (yes, no)
- Which businesses were named, in order?
- Which sources were cited? (URLs or site names)
- Did the answer contain a factual error about you or your market?
The cited sources column is the most valuable part of the audit. In our case, Clutch, Expertise.com, and DesignRush accounted for the majority of citations for agency recommendations. Competitors also appeared through their own city-specific service pages, not through their homepages.
Step 4: Score the results
Three numbers tell the story:
- Mention rate: the percentage of the 150 answers that named you. Zero to 5 percent is typical for a first audit.
- Competitor mention rate: who appears most often. This is your real competitive set for AI search, which may differ from your Google competitors.
- Source concentration: how many distinct sources account for 80 percent of citations. If it is five or fewer, you know exactly where to work.
Step 5: Act on what you found
If directories dominate the citations: claim, complete, and get reviews on the ones cited. An incomplete profile with no reviews is invisible to AI.
If competitors’ own pages are cited: look at what those pages have that yours do not. Usually it is a page per city or service with a direct answer at the top, specific pricing or process detail, and structured data.
If the AI has wrong facts about you: old addresses, closed offices, wrong phone numbers. Find the source (often a data aggregator) and correct it. We found two aggregators still publishing office addresses we had closed years earlier.
If you are absent from brand queries: the AI does not have enough information to describe you. That points to schema markup, an About page that states plainly what you do and where, and consistent business information across the web. The structured data side is covered in Schema markup for AI search.
Step 6: Re-run quarterly
AI answers change constantly. Repeat the same 30 queries every quarter, keep the spreadsheet, and track your mention rate over time. Improvement usually shows up within two quarters of fixing citations and adding structured, answer-first content.
Frequently asked questions
Do I need paid tools to run this audit?
No. All five platforms are free to query. Paid AI visibility tools automate the process and add volume, but the manual method gives you the same insight for a first baseline.
Why do different AI platforms give different answers?
Each uses different underlying search data and weights sources differently. Perplexity and ChatGPT lean on live web results; Gemini and AI Overviews lean on Google’s index. A business can appear in one and be absent from another.
How long until AI assistants start recommending my business?
Two to six months after fixing the underlying sources is typical. Directory updates and review growth show up faster than content changes.
Should I ask the AI to recommend my business?
No. Prompting an AI about yourself has no effect on what it tells other users.
Is AI search replacing Google?
Not yet, but it is taking the top of the funnel. Businesses that appear in both are winning the customers who ask AI first and confirm on Google second.
