A 5-Minute Test That Shows Whether ChatGPT Recommends a Business

A restaurant owner in Brooklyn ran a simple experiment last month. She opened ChatGPT on her phone, typed “best Italian restaurant in Park Slope,” and hit send.

Her restaurant wasn’t in the answer. Three competitors were.

She tried again with different wording. “Italian food near me in Park Slope.” “Where to eat pasta in Brooklyn.” Five tries, five answers. Her restaurant showed up once, in fourth position. One of her competitors appeared every single time, always in the top two.

She’d been running the place for twelve years. Great Google reviews, a loyal customer base, steady foot traffic. But a growing number of people don’t Google restaurants anymore. They ask ChatGPT. And ChatGPT had a clear favorite in her neighborhood, and it wasn’t her.

The Test Takes Five Minutes

Any business owner can do a basic version of what she did. No technical skills needed. Just a phone and a few minutes.

Step one. Open ChatGPT (or Gemini, or both). Don’t sign into a business account or anything special. Just open it the way a normal customer would.

Step two. Ask the kind of question a potential customer would ask. Not the business name. The category and location. “Best pizza in Austin.” “Good dentist in downtown Denver.” “Where to get a haircut in Silver Lake.” Think about what someone who doesn’t know the business yet would type.

Step three. Look at the response. Which businesses appear? In what order? How are they described?

Step four. Ask the same thing three or four more times, with slightly different wording. “Top rated pizza in Austin” instead of “best pizza in Austin.” “Recommend a dentist near downtown Denver.” The answers might change. They might not. Both outcomes are interesting.

Step five. Try the same queries on a different platform. If the first test was on ChatGPT, try Gemini. The results are often surprisingly different.

What to Look For

The test reveals three possible scenarios.

The business shows up consistently. It appears in most or all responses, in a decent position. Good news. But worth tracking, because these recommendations can shift over time without warning.

The business shows up sometimes. It appears in some queries but not others, or only on one platform. This is the most common scenario and the trickiest. It means the business is on the edge of being recommended. Small changes in how someone phrases a question can make it appear or disappear.

The business doesn’t show up at all. It’s not in any of the responses, on any platform, regardless of phrasing. For the growing number of people who ask ChatGPT before making a decision, this business simply doesn’t exist.

Why the Results Look Different Every Time

Something that catches people off guard during this test is that the same question can produce different answers on different days, or even minutes apart. ChatGPT and Gemini aren’t static directories. They generate fresh responses each time, and the results can vary.

This is actually one of the most important things to understand. Unlike Google, where a business holds a relatively stable ranking, ChatGPT recommendations are fluid. A business that showed up on Monday might not show up on Wednesday. Which makes a single manual check useful as a starting point, but unreliable as a complete picture.

In our research on pizza recommendations in NYC, we found that some businesses hit 100% mention rate on one platform and 0% on another. Mama’s Too appeared in every single Gemini response and in zero ChatGPT responses. That’s not a difference that one quick test would reliably catch.

What a Manual Check Can’t Tell

The five-minute test is a good start. It answers the most basic question: “Does ChatGPT even know this business exists?” But there are things it can’t do.

It can’t show patterns. Running four or five queries gives a snapshot. Not a trend. A business might appear 80% of the time across a large sample but happen to miss the five specific queries an owner tried. Or the opposite.

It can’t track changes. Maybe ChatGPT recommended the business last month but stopped this week. Without regular monitoring, there’s no way to know when things shift or why.

It can’t compare across platforms at scale. Checking ChatGPT and Gemini manually is doable. Adding Perplexity, Copilot, and whatever new platforms launch next month? That becomes a part-time job.

It can’t normalize names. Our NYC pizza research found the same business listed under four different spelling variations. A manual search for the exact business name might miss mentions that use a slightly different version.

It can’t show what competitors are doing. The test reveals who else shows up. But it doesn’t explain why they show up, how often, or whether their position is improving or declining over time.

What Actually Drives These Recommendations

This is still an evolving picture, but some patterns are becoming clearer.

ChatGPT and Gemini pull from a combination of sources when forming recommendations. Review sites, editorial content, blog posts, social media mentions, curated lists, local directories. The weight each platform gives to each source is different, which is why their recommendations often disagree.

A few things seem to matter consistently. Volume and freshness of reviews. Mentions in editorial and “best of” lists. Consistent business information across directories. Active online presence beyond just a Google Business profile.

But there’s no public algorithm to reverse-engineer. No one outside of OpenAI and Google knows exactly how these recommendations get assembled. What businesses can do is make sure the information that exists about them online is accurate, detailed, and spread across multiple sources. And then actually measure whether it’s working.

Beyond the Five-Minute Check

The manual test is where every business owner should start. It costs nothing and takes five minutes. Even just knowing the answer to “what does ChatGPT say about my business” is more than most competitors have bothered to find out.

But for anyone who wants the full picture, Reachd.ai continuously monitoring how a business appears across ChatGPT, Gemini, Perplexity, and other search platforms, catching shifts in recommendations, name variations, and competitive changes that manual spot-checks miss. It also generates plain-language action items, which are specific, prioritized steps any business owner can follow to improve their AI visibility, no technical background required.

The five-minute test opens the door. What’s behind it is usually worth a closer look.