How to Check If Your Business Shows Up in AI Search

To check whether your business shows up in AI search, ask each engine the questions your customers actually ask, in a fresh session with personalization and location history switched off, then record which businesses get named and which sources get cited. Run the same prompts in ChatGPT, Google AI Overviews, Perplexity and Gemini, because each one pulls from different sources and the answers rarely match.

Most business owners have never done this. They assume that ranking on Google means being mentioned by AI. Those are now two different results, produced by two different systems, and it is entirely normal to rank on page one of Google and be completely absent from the AI answer above it.

This guide is the diagnostic. It takes about thirty minutes, costs nothing, and tells you exactly where you stand before you spend money on fixing it.

Why your Google ranking does not tell you this

Traditional search returns a list of links and lets the user choose. AI search returns a decision. When someone asks an assistant to recommend a roofer in Doylestown, they get two or three names, not ten blue links.

That changes what “visibility” means. Three differences matter:

• Being ranked is not being cited. AI systems synthesize an answer from sources they trust and often name businesses that are not the top-ranked result for that query.
• The answer changes by phrasing. “Best roofer near me” and “who should I hire to replace a roof in Bucks County” can return completely different businesses.
• There is no rank position to check. You are either named or you are not. There is no page two.

So the only reliable method is to go and look, systematically, with a fixed set of prompts you can repeat later to measure change.

Step 1: Write the questions your customers actually ask

Do not test your brand name first. Searching your own business name will almost always return something, and it tells you nothing about whether you get discovered by people who do not know you exist.

Build a list of ten to fifteen prompts in three groups:

Prompt type	Example	What it tests
Discovery	“Who are the best window treatment installers in Bucks County, PA?”	Whether you get named at all to a stranger
Comparison	“Should I hire a local marketing agency or a national one for a service business?”	Whether you are used as an example or authority
Problem-first	“My patio is unusable in summer because of the sun. What are my options?”	Whether you appear before the customer knows your product category
Branded	“What does [your business] do, and where are they located?”	Whether the AI has your facts right

 

The problem-first prompts are the ones most businesses skip, and they are where the earliest customers are. Someone describing a problem has not yet decided what to buy.

Step 2: Test in a clean session

This is the step that ruins most people’s results. If you are signed in, in your own office, on the network where you have visited your own website a thousand times, the assistant will show you your business and you will conclude everything is fine.

For each engine:

• Sign out, or use a private or incognito window.
• In ChatGPT, disable memory and personalization, or use a temporary chat.
• Do not include your business name in the prompt.
• Note the location the tool assumes. If it is guessing your city from your connection, say the target location explicitly instead: “in Cherry Hill, NJ” rather than “near me.”
• Run each prompt twice. AI answers are not deterministic, and a business that appears once out of two runs is in a very different position from one that appears every time.

Step 3: Check all four engines, because they disagree

Engine	What it draws on	What to record
Google AI Overviews	Google’s index, plus strong local signals from Google Business Profile and Maps	Are you named? Which pages are linked in the sources panel?
ChatGPT	Its training data plus live web results when browsing	Are you named, and does it cite a URL or state it from memory?
Perplexity	Live web retrieval with visible citations on every claim	Which of your pages are cited, and which competitor pages beat them
Gemini	Google’s index and Google surfaces	Whether local and Maps data pulls you in where the web content does not

Perplexity is the most useful diagnostic of the four, because it shows its sources for every sentence. If a competitor is cited and you are not, you can open the exact page that won and compare it to yours. That comparison is worth more than any visibility score.

Step 4: Record it in a way you can repeat

A one-time look is an anecdote. Put it in a simple sheet so it becomes a baseline:

Prompt	Engine	Date	Named? (Y/N)	Position in the answer	Your pages cited	Competitors named

Re-run the same prompts monthly. The number that matters is not a score out of 100. It is the share of your prompts where you get named, tracked over time, and whether the competitors named alongside you are changing.

Be careful with third-party “AI visibility” tools that report a single percentage. They are useful for trend direction, but they run their own prompt sets, which may have nothing to do with how your customers speak. Your own prompt list is the more honest measure.

How to read your results

Once you have run the prompts, you will land in one of four situations. Each has a different fix.

You are never named, and no page of yours is ever cited. The engines do not have enough about you to work with. This is the most common result for local service businesses and it is usually a content problem: your site describes what you sell but never answers the questions customers ask. Start with the questions, not the services.

You are named on branded prompts only. The engines know you exist but do not consider you an answer to the category question. You have an authority and evidence gap. Case studies, specifics, named results and third-party mentions are what close it.

You are named, but the details are wrong. Old address, wrong hours, a service you dropped two years ago, a phone number from a previous listing. This is the fastest problem to fix and the most damaging to leave. It usually traces back to inconsistent information across your site, your Google Business Profile and directory listings. See our guide to optimizing your Google Business Profile for AI Overviews.

You are named alongside three competitors, consistently. This is a good position. The work now is displacement: understand which pages get cited for your competitors and produce a better answer to those specific questions.

The five things that most often make a local business invisible to AI

These are the patterns we see repeatedly when we run this audit for service businesses.

1. No page answers a question. Service pages that describe an offering, with no page anywhere that directly answers “how much does this cost,” “how long does it take,” or “what is the difference between X and Y.” AI systems quote answers. If you have not written one, there is nothing to quote.
2. The answer is buried. The information exists, but it is in paragraph nine, or inside an accordion or tab that only opens on click. Content hidden behind JavaScript interactions is frequently not read. Put the direct answer in the first hundred words, in plain text.
3. Inconsistent business facts. Your name, address and phone number differ between your website footer, your Google Business Profile and old directory listings. Conflicting facts reduce confidence, and low confidence means the engine picks someone else.
4. No evidence. Claims with nothing behind them read as marketing. Named results, real numbers, dated case studies and specifics are what get quoted. Vague superlatives never are.
5. No structured data. Schema markup does not force a citation, but it removes ambiguity about what your page is, who wrote it and what questions it answers. On a page with a genuine FAQ section, FAQPage markup makes those question-and-answer pairs machine-readable rather than something the engine has to infer.

Our longer piece on why some content gets cited by AI goes deeper into the patterns behind points one, two and four.

How this connects to SEO, AEO and GEO

These are not competing strategies, and you do not choose between them.

• SEO gets your pages indexed, crawlable and ranking. It is the foundation, because a page that cannot be found or read cannot be cited.
• AEO, answer engine optimization, structures your content so it can be extracted as a direct answer.
• GEO, generative engine optimization, builds the authority and consistency that makes a generative engine willing to name you.

If you want the full explanation of how the three fit together, we cover it in our guide to answer engine optimization. The short version: AEO and GEO are not a replacement for SEO. They are what you do after the SEO fundamentals are in place, and they are increasingly where the earliest, highest-intent customers are.

For local service businesses there is a fourth piece. AI Overviews and Gemini lean heavily on Google Business Profile and Maps for anything with local intent, which means your local SEO and your AI visibility are the same project, not two.

A realistic timeline

Nobody can promise you a citation, and you should be skeptical of anyone who does. What we can describe is the usual sequence.

Weeks 1 to 2. Fix factual inconsistencies across your site, Google Business Profile and major listings. This is the fastest-moving work and it is the reason wrong-detail problems resolve first.

Weeks 2 to 6. Publish direct answers to the questions from your prompt list. One page per question cluster, with the answer stated plainly at the top.

Weeks 6 to 12. Add evidence. Case studies with real numbers, dated. Then re-run your prompt sheet and compare against your baseline.

Ongoing. Re-run monthly. AI answers shift as models update and as competitors publish. A citation you earned in March is not permanent.

The honest caveat: this varies enormously by industry and by how much content already exists in your category. A niche local trade with few competitors moves faster than a saturated market.

Want the audit done for you?

We run this exact diagnostic for service businesses: a prompt set built around how your customers actually speak, tested across all four engines, with a baseline sheet and a prioritized fix list.

Some of what that work has produced for clients: a Philadelphia venue at 3.52x ROAS, a window company generating $1,413,465 in revenue from $199,920 in spend, and a home improvement client running $148,720 signed from $31,984 in ad spend.

Evolve Demand is a Philadelphia-based performance marketing agency working with 20+ service businesses across home improvement, entertainment, healthcare and local trades. We are rated 5.0 on Google. Our AEO and GEO services for small businesses are the work behind these numbers.

Request a free AI visibility audit · See our work · contact@evolvedemand.com