Most local business owners have no idea whether ChatGPT, Perplexity, or Google AI Mode is recommending them. Not because they don't care. They've read enough headlines to know the shift is happening. But no one has told them exactly what to type and where to look.
This post does that. In about 20 minutes, you'll have a real baseline: which AI engines know about your business, which ones are sending customers to your competitors instead, and what the gaps look like. That baseline is the starting point for everything else.
Why this matters before you do anything else
Both SEO and GEO strategies produce the same temptation: jump straight to fixing things. Add schema. Collect reviews. Optimize the Business Profile. All good moves. But if you do them without a baseline, you have no way to know whether they worked. You're flying blind in both directions.
Running the visibility check first takes 20 minutes. It tells you whether you have an AI citation problem at all, which engines have the worst picture of your business, and which competitors are getting cited instead of you. That shapes everything you prioritize next.
What you're actually measuring
When someone types a local query into ChatGPT, Perplexity, Claude, or Google AI Mode, the engine synthesizes an answer from sources it has indexed or been trained on. Sometimes it names businesses. Sometimes it names your competitors. Sometimes it doesn't name anyone and gives generic advice instead.
Your goal is to find out what happens when a potential customer asks a question you should be the answer to. Not whether you rank on page one of Google — that's a separate check. Specifically whether the AI names you, recommends you, or completely ignores you when the query is squarely in your wheelhouse.
The setup: pick your five test queries
Before you open any AI engine, spend five minutes writing five queries that a real customer would realistically ask. Not your business name — they already found you if they know your name. The queries you want are the ones customers type when they're looking for a solution, not a specific business.
Write these down. You'll run the same five queries across all five engines so the comparison is apples to apples.
The check: five engines, one at a time
Open each of the following and run your five queries. For each query, note: were you named? Were your competitors named? Did the engine give a useful local recommendation at all?
ChatGPT
chatgpt.com
Type your query directly into the chat. ChatGPT will usually respond with a short explanation followed by a list of businesses or a general answer with caveats about not having real-time data.
If ChatGPT says it can't recommend specific local businesses, that's actually common — its training data has a cutoff and local listings aren't its strength. Note this and move on; it doesn't mean you've done something wrong.
If it does name businesses, screenshot it. If it hedges heavily (“I'd recommend checking Google Maps for current options”), try a slightly different phrasing — more specific, more conversational.
Perplexity
perplexity.ai
Perplexity is the engine most aggressively answering local queries with current web data. It does real-time web searches and cites sources, so results are more like a live research summary than a trained response.
Look at what it cites and whether your website, Google Business Profile, or any directory listing appears. If your competitors show up as cited sources and you don't, that's important data: Perplexity is finding them on the open web and not finding you. Also pay attention to how Perplexity frames the answer — review volume vs. service descriptions tells you which gap to close first.
Claude
claude.ai
Claude (the AI, not Seen) is more conversational and less likely to name specific businesses than Perplexity, but it increasingly has web access and will pull current data on local queries. Claude tends to add caveats about verifying with Google Maps — that's normal. What matters is whether, when it does name businesses, yours is among them.
Google AI Mode
google.com, if available
This is the highest-leverage check. Google AI Mode sits at the top of the search results page for many local queries and influences more decisions than any other AI surface right now — because it's where your customers already are.
Search your five queries. If you see an AI-generated summary at the top — a conversational block that synthesizes an answer before the traditional results — that's AI Mode. Screenshot every one. Note whether you're in the summary. If you're ranking #1 organically below the fold but not in the AI summary above it, you now know exactly what the problem is.
Grok
grok.com or via X
Grok is newer to local search but worth a check, especially if your business has any social presence on X. It's the engine most likely to surface businesses that are actively discussed on social platforms, so if you have an engaged audience there, that can translate into citations here.
Recording what you find
After running all five queries across all five engines, you should have a grid something like this in your notes:
| Query | ChatGPT | Perplexity | Claude | AI Mode | Grok |
|---|---|---|---|---|---|
| “your query 1” | Not named | Competitor A | Not named | Competitor B | Not named |
| “your query 2” | Named ✓ | Named ✓ | N/A | Not in summary | Not named |
The pattern tells you where to focus. If you're invisible across all five engines on all five queries, GEO is the whole game and the gaps are likely foundational: incomplete GBP, thin schema, no third-party mentions. If you're showing up in two engines but not three, you have partial signals and the fix is usually filling the gaps in structured data and directory consistency.
The two numbers that matter most
Your citation rate
Out of 25 total engine-query combinations (5 queries × 5 engines), how many returned a citation of your business? If you're at 3 out of 25, your baseline citation rate is 12%. That's the number you'll want to move.
Competitor citation rate
Pick your two main competitors and run the same queries for them. If they're at 14 out of 25 and you're at 3 out of 25, the gap is specific and measurable. That's more useful than a vague sense that you're “not showing up in AI.”
What the results typically look like
The blank slate: AI engines either say they can't recommend specific businesses or give highly generic answers. This usually means the business has thin structured data and minimal third-party presence. The AI has nothing useful to say about it. The fix starts with GBP completeness, schema markup, and getting listed in the relevant industry directories.
The competitor showcase: The engines consistently name two or three well-known competitors and omit you despite similar quality and reviews. This usually means the competitors have been around longer, have more review volume, or have more consistent data across platforms. The gap is closable. It's a signal consistency and volume problem, not a fundamental visibility problem.
The partial picture: The engines know you exist but describe you incorrectly — wrong hours, outdated services, missing specialties. This is actually a common and fixable situation. It means AI engines have found your business in some sources but those sources are incomplete or out of date. The fix is a NAP audit and a structured data refresh.
The strong start: You're showing up in two or three engines on most queries. This means your foundational signals are working. The next step is earning third-party mentions and building FAQ and service content that gives engines more structured material to cite.
After the check
The visibility check gives you a baseline. What you do with it depends on what you found, but the next moves are usually one of these three:
If you're invisible across most engines, start with Google Business Profile completeness, then schema markup on your top service pages, then a NAP audit across the five major directories. Those three moves cover most of the foundational gap.
If you're partially visible — you show up in one or two engines but not others — the work is about consistency. Make sure your service descriptions, hours, and categories say the same thing everywhere. Add FAQ content to your site. Build out your author and team bios so engines can verify who you are.
If you're showing up but losing to competitors on specific queries, this is a content and mention-building problem. You need pages that directly address the queries you're losing, and you need third-party sources — local press, industry directories, community mentions — to corroborate what you're saying about yourself.
Run the check again in 60 days. AI engines re-index on their own cadence — citations often take 6 to 8 weeks to appear after fixes are applied. The only way to know whether your work is moving the number is to have a number to compare against.
The 20-minute summary
- 1Write five queries a real customer would ask before they know your name.
- 2Run them in ChatGPT, Perplexity, Claude, Google AI Mode, and Grok.
- 3Screenshot every result that names a business.
- 4Calculate your citation rate and your top competitors' citation rates.
- 5Identify the pattern: blank slate, competitor showcase, partial picture, or strong start.
- 6Pick the right first move based on the pattern.
- 7Set a calendar reminder to run the check again in 60 days.
The check costs 20 minutes. The gap it reveals costs customers every week it goes unaddressed.
Frequently asked questions
How long does the visibility check take?
About 20 minutes for all five engines. Set aside a bit more time the first time because you'll be noting patterns and writing down competitor citations. Once you have a routine, subsequent checks go faster.
What if ChatGPT says it doesn't have real-time local data?
That's common and not a reflection of anything wrong on your end. Note the response and move on to Perplexity and Google AI Mode, which are better at current local data. Use ChatGPT in your baseline grid as a data point, not a verdict.
How often should I run this check?
Every 60 days is a practical cadence. AI engines re-index on their own schedule, and changes to your GBP, schema, or content typically take 4 to 8 weeks to show up as citation improvements. Running every 60 days gives changes enough time to take effect before you measure again.
Which engine matters most?
Google AI Mode has the highest impact because it sits at the top of search results where your customers already are. If you only check one engine, check Google AI Mode. But running all five gives you a complete picture since different engines weight different signals.
Skip the spreadsheet
Let Seen run this check automatically
Across all five engines, every week, against the queries that matter most to your business. You get your citation rate, your competitors' citation rates, and the exact gaps — without opening a single chat window.
Start your free audit