How to know if ChatGPT recommends your business
Your buyers are not asking Google anymore. Research from early adopters shows that for high-consideration B2B and consumer purchases, more than half of buyers now ask ChatGPT, Claude, or Perplexity for a recommendation before they ever click a search result. The AI answers with names. Some of those names are your competitors. Some are companies you have never heard of. The question every founder and marketing operator should be able to answer right now is simple: is your business one of the names AI is giving?
This guide gives you the 5-minute test, explains why one test is not enough, and walks through the five signals that actually change whether AI engines cite you. By the end you will know exactly how to find out where you stand and the specific work to do if the answer is "not yet."
The 5-minute test you can run right now
Open ChatGPT and type the question your buyer would actually type. The exact phrasing matters. Buyers do not search for product names. They describe the job they need done.
If you sell competitive intelligence software, the buyer types something like "what is the best competitive intelligence software for a marketing agency". If you run a dental practice in Toronto, the buyer types "who is the best dentist in downtown Toronto for invisalign". Use the buyer's words, not yours.
Run that exact prompt in four engines: ChatGPT, Claude, Perplexity, and Google's AI Overviews (you get those automatically when you search Google for the same query). Note which businesses each engine names. If your business is not on any of the four lists, you have your answer.
Why running the test once is not enough
AI engines are non-deterministic. The same prompt at 9am and 9pm can produce different lists. The same prompt from a New York IP and a Toronto IP can produce different lists. Models update their training data; some change weekly. A single test tells you what one model said one time.
What you actually want to know is the citation rate: across N runs, across the four major engines, across multiple plausible buyer phrasings, how often does your business get named, and at what rank? That number is the leading indicator of whether AI is sending you customers.
Practical advice: run the same query at least three times across each engine. Then run two or three rephrasings of the same buyer intent. That gives you 9 to 12 data points per engine. If you appear in fewer than 30% of the runs, you have an AI visibility gap that is costing you leads today.
What signals tell ChatGPT to cite you
AI engines do not crawl the web the way Google did. They build an internal model of which businesses are credible authorities for a given category, then they cite from that model. The signals they use are largely structured data and entity verification, not backlinks and keyword density.
The five signals that move citations the most:
1. Organization schema with sameAs entity links
A JSON-LD Organization schema on every page of your site, with a sameAs array linking to your LinkedIn company page, X profile, Google Business listing, and any industry directories. AI engines use these as identity-confirmation signals. Without them, your business looks like an unverified entity. With them, your name resolves to a confident, structured record.
2. FAQ content matching buyer language
FAQ sections written in question-and-answer format using the exact phrasing buyers use. Each Q&A pair is a direct extractable signal. AI engines lift FAQ answers verbatim into their responses. If your service page does not have a FAQ, you are leaving the most extractable content type off the table.
3. Author and Person schema (E-E-A-T)
For service businesses and consultants, AI engines weight author authority heavily. A Person schema for the founder or expert behind the business, with knowsAbout claims and a sameAs link to LinkedIn, is one of the highest-leverage additions you can make.
4. Citation-ready content depth
Pages under 500 words are rarely cited by AI engines because they do not contain a complete enough answer to lift. Service pages and category-explainer pages of 1,200 to 2,000 words, written in clear sub-sections with question-shaped H2s, get cited consistently. The audit you can run on yourself: do your pages answer the buyer's question fully, or do they pitch your service and stop?
5. External authority links
Outbound links from your content to authoritative sources (Wikipedia, schema.org, government sites, industry research) signal that you are a credible part of an information network rather than an isolated marketing site. Counterintuitively, linking out to authority is one of the strongest AI-trust signals available.
See your AI visibility for free
Run the test we just described, automated across all four AI engines, in 90 seconds. No credit card.
Run free scan →How long does it take?
Implementing the five signals above is fast. A skilled developer can ship the schema, the FAQ structure, and the author markup in a few hours. The slow part is the AI engines re-crawling your site, re-processing it, and updating their internal models. The realistic timeline from "I shipped the changes" to "ChatGPT consistently names me" is 8 to 12 weeks.
That timeline is not negotiable through code. AI training and re-ranking cycles run on their own clock. The companies that win the 2026 search shift are the ones that ship the signal work this quarter, accept the 8-to-12-week wait, and let the citations compound from there. The companies that wait until next year will be 2 to 3 quarters behind a competitor who shipped in May.
What about Google? Is SEO still the same job?
SEO and AEO overlap on the technical foundation (canonical URLs, semantic HTML, fast page loads, hreflang for international audiences) but diverge on what gets you cited. SEO rewards backlinks, keyword presence, and page authority. AEO rewards structured data, entity verification, and citation-shaped content. A page that ranks first on Google can be invisible to ChatGPT. A page that gets cited by ChatGPT can be unranked on Google.
For most businesses, the answer is to do the SEO foundation work AND add the AEO signals on top. The SEO foundation also helps the AEO score, because AI engines use search engine signals as one input. The opposite is not as true; AEO signals help your search ranking only modestly.
What to do this week
- Run the 5-minute test. Use ChatGPT, Claude, Perplexity, and Google AI Overviews. Note which engines name you, which name competitors, and what the gap looks like.
- Audit your structured data. Open one page on your site, view source, search for
application/ld+json. If you find none, that is your starting work. - Add Organization schema. Even a minimal Organization block with name, URL, logo, and a sameAs array is a meaningful upgrade from nothing.
- Write one FAQ page. Pick the five questions buyers actually ask you and write 200-word answers using the buyer's exact phrasing.
- Set a 12-week check-in. Re-run the same test 12 weeks from today. Track whether your citation rate moved.
The honest version
You do not need a $499/month AI visibility tool to run the 5-minute test. You can do that today, for free, in your browser. Where tools like Apex Radar earn their keep is the part that comes after: continuously running the citation tests across engines and locations, tracking the rate over time, comparing it to specific competitors, and surfacing exactly which signals to add when the rate stalls.
If you are pre-launch or under 100 employees, you might not need that yet. Run the manual test, ship the five signals, wait 12 weeks. If you do need ongoing monitoring, that is what we built Apex Radar for, and the free baseline scan gets you the same data the manual test would, in 90 seconds, automated across all four engines.
Either way, the most expensive thing you can do is not run the test.