Asset 20 8 2
Does AI recommend your business? Run the free check →

Join 15,000 business owners, marketers and entrepreneurs. The Sunday newsletter you'll be annoyed only arrives once a week.

Article

How to Check Whether ChatGPT Is Recommending Your Business

The short version: open a fresh, logged-out ChatGPT session and ask the questions a customer would type, not questions with your brand name in them. Do it across ten to fifteen prompts, across more than one AI tool, and repeat it weekly, because the answer changes from session to session. One good result doesn’t mean you’re “in,” and one bad result doesn’t mean you’re out either.

Why most people check this wrong

I did this test badly the first time I tried it, so I’ll save you the embarrassment. I typed “is Lilach Bullock a good marketing consultant” into ChatGPT, got a lovely answer, took a screenshot, and told myself I’d “cracked” AI visibility. Three days later I typed almost the same question and got a completely different, blander answer that didn’t mention me at all. Nothing had changed on my website. Nothing had changed about my business. What changed was the prompt, the session, and the mood of the model that day.

That’s the bit nobody tells you when they write about this topic: checking your own brand name isn’t checking whether ChatGPT recommends you. It’s checking whether ChatGPT can find you when you’ve already told it who you are. Real customers don’t do that. A real customer in Maidstone doesn’t type “tell me about Bloom Bakery.” They type “best bakery for a birthday cake near Maidstone” or “who does good custom cakes in Kent.” That’s the question that matters, and it’s the one most business owners never test.

The proper way to run the test

Here’s the exact process I now run for myself and for clients, roughly once a fortnight.

  • Step 1: Go incognito. Open a private browser window and use ChatGPT logged out, or logged into a fresh, unused account. If you check while logged into your normal account, ChatGPT may be using memory of your past chats, which skews results in a way a stranger’s browser won’t.
  • Step 2: Write ten customer-style prompts, not brand prompts. Think in terms of the problem, not the company. For a plumbing client in Sevenoaks, our list included “emergency plumber near Sevenoaks,” “who fixes boiler leaks in Kent,” “reliable local plumber good reviews,” and “plumber that does weekend call outs Sevenoaks.” None of those mention his business name. That’s deliberate.
  • Step 3: Run each prompt three times. Not once. Three times, spaced a few minutes apart. Note whether your business shows up, in what position, and whether the description is accurate.
  • Step 4: Test more than ChatGPT. Run the same prompts through Perplexity, Google’s AI Overviews, and Microsoft Copilot. They pull from different sources and give different answers. I’ve had clients show up consistently in Perplexity, which cites live web sources, and disappear completely in ChatGPT’s default mode, which leans more on training data and general web knowledge unless browsing is switched on.
  • Step 5: Log it. Build a simple spreadsheet: date, prompt, tool used, mentioned yes/no, position in the answer, accuracy of the description. This is the step people skip and it’s the one that shows you a trend rather than a snapshot.
  • Step 6: Check what it says about you when it does mention you. Wrong opening hours, an old address, a phone number for a business you sold five years ago, all of this happens, and it’s arguably worse than not being mentioned at all. I’ve written before about the AI chatbot problem nobody warns small businesses about, and inaccurate recommendations sit right in the middle of it.

A real example, with real numbers

When I ran this test for a client who runs a small accountancy firm near Tunbridge Wells, I asked the same prompt, “best small business accountant Tunbridge Wells,” twenty separate times across four days, alternating between a fresh browser and a fresh account each time. Her firm appeared in eleven of the twenty answers. Two competitors appeared more consistently, at sixteen and fourteen mentions respectively. One competitor who has been trading for less than two years, with a fairly basic website, beat her firm of eleven years simply because they had recent Google reviews with specific keywords in them (“great with self assessment,” “sorted my VAT in a week”) and a Companies House-linked profile that made their existence easy to verify.

Eleven out of twenty isn’t a fail. It also isn’t a pass you can rely on. That’s the uncomfortable bit. There is no fixed score sitting somewhere that says “ChatGPT recommends you: yes or no.” It’s probabilistic, it shifts with model updates, and a result from Tuesday tells you almost nothing about what happens on Friday. If you’ve read pieces claiming they’ve found “the ChatGPT ranking algorithm” for local businesses, be sceptical. There isn’t one to find in the way there’s a Google ranking to check.

What moves the needle, based on what changed for her

Once we had the log, we didn’t chase the checking, we chased the signals that seemed to correlate with better mentions:

  • Recent Google reviews, ideally with specific service terms in them, not just “great service, five stars.”
  • A Google Business Profile that was fully filled in, categorised correctly, and updated in the last few months.
  • Being named on other people’s websites, not just her own; local business directories, a mention in a Kent business Facebook group post, a guest spot on a local radio station’s website.
  • Clear, structured information on her own site: services listed plainly, an actual address, actual opening hours, not buried in a PDF.

Three months later, the same prompt run twenty times returned her firm sixteen times. Nothing about her business had changed. Her reviews had grown from 34 to 61, and she’d picked up two directory listings and a short write-up on a local Chamber of Commerce page. That’s the whole trick, and it’s less exciting than people want it to be. ChatGPT doesn’t have a special back door for AI visibility. It’s reading the same web everyone else reads, plus a bit of training data, plus whatever Bing or its search partner surfaces at that moment.

The part that’s uncomfortable

Here’s the thing most posts on this topic won’t say out loud: for a huge number of small businesses, especially very local, very small ones, ChatGPT is never going to recommend you with any consistency, no matter what you do, because there simply isn’t enough written about you anywhere on the internet for it to have anything to work with. If you’ve got twelve reviews, no press mentions, no directory listings beyond Google, and a website with three pages, you are, from the model’s point of view, close to invisible. That’s not a technical problem you fix with a plugin. That’s a “you need more of the internet to know you exist” problem, and it takes months, not a weekend.

I’ve watched business owners spend hours running prompt after prompt, tweaking their meta descriptions, adding “AI-friendly” FAQ schema to their homepage, convinced that’s what will flip the switch. Sometimes it helps a little. It rarely does what people hope, because the underlying issue is that the raw material AI models pull from about that business is thin. You can’t optimise your way out of not existing much online. You have to go and exist more: get quoted, get reviewed, get listed, get mentioned by other people, in other places, in your own words matching the words customers use.

This is also why I’d be careful of anyone selling a one-off “ChatGPT visibility audit” as if it’s a fixed, one-time fix. If you want proper help with this, look for someone who treats it as ongoing work, checking, tracking, building up mentions over months, not a single report. That’s the difference between what a decent AI consultant for small business does and what a quick audit PDF pretends to do.

Watch what happens when you don’t check

The flip side is worth saying plainly. If you never run this test, your competitor might be showing up in your place and you’d have no idea. I wrote a whole piece on this exact scenario, why ChatGPT might already be recommending your competitor, after finding a client’s biggest rival being named first, by name, in answer to a prompt the client had never bothered to test. He’d assumed that because his SEO was solid and he ranked well on Google, he’d be fine everywhere else. He wasn’t. Google ranking and AI recommendation overlap, but they’re not the same list.

If you want a fuller walkthrough of the fix side of this, not just the checking side, my companion piece on whether ChatGPT is recommending your small business, and how to fix it goes into the practical work: what to update on your Google profile, how to get review text that helps, and which directories are worth the ten minutes it takes to list yourself. Worth reading side by side with this one.

How often to check

Weekly is overkill for most small businesses and will just make you anxious about noise. Monthly is sensible. Quarterly is fine if you’re busy and your industry isn’t changing fast. What matters more than frequency is consistency of method: same ten prompts, same rough process, logged the same way, so month three is comparable to month one. Random one-off checks whenever you remember are close to useless because you’ve got nothing to compare them against.

Also worth knowing: the AI landscape itself keeps shifting under all of this. Model updates change how these answers get generated almost every month now, which is part of why I cover what’s changed each week in my AI news roundup for small business, because a change to how ChatGPT handles browsing or local search can quietly move your results without you touching a single thing on your website.

A quick sanity check before you panic

Before you conclude AI hates your business, run the same test on three known, well-established competitors in your exact area. If they’re barely showing up either, the issue is likely industry-wide or category-wide, not personal. Some categories, plumbers, dentists, electricians, tend to surface well because there’s a lot of review and directory data. Others, very niche B2B services for example, barely get mentioned for anyone, because the training data and web content for that niche is thin across the board. Context matters more than your ego wants it to.

Frequently asked questions

How do I know if ChatGPT is recommending my business?

Test it with customer-style prompts, not your business name, in a fresh logged-out session, running each prompt two or three times. Log whether you’re mentioned, in what position, and whether the details are correct, then repeat monthly so you can see a trend rather than a single snapshot.

Why does ChatGPT give different answers to the same question?

ChatGPT’s answers are probabilistic, meaning the model generates a response fresh each time rather than pulling from a fixed, stored ranking. Session settings, whether browsing is switched on, and recent model updates all affect the answer, so the same prompt can return different results within the same week.

Does ChatGPT use Google rankings to decide who to recommend?

Not directly, though there’s overlap. ChatGPT’s answers are shaped by its training data plus, in some modes, live web browsing that often draws on similar signals to search engines: reviews, directory listings, and how often a business is mentioned elsewhere online. Ranking well on Google helps but doesn’t guarantee an AI mention.

What’s the fastest way to improve my chances of being recommended?

Get more recent, specific Google reviews, fully complete your Google Business Profile, and get listed or mentioned on other websites you don’t control, like local directories or press. These build up the raw material AI tools pull from, and they tend to matter more than anything you can do purely on your own website.

Useful references

Related reading: Why Your LinkedIn Account Got Restricted (and How to Fix It) and How to Schedule an Entire Photo Album on Facebook in Advance.

Want the complete version? Read where I break down AI marketing.

You can free AI Visibility Checker in about a minute: it queries five engines and scores whether they mention you, then audits the page for schema, llms.txt and answer-first structure.

Published and maintained by the Lilach Bullock team, covering marketing, AI and business growth.
Your buyers are asking AI who to use. Does it say you?

See for free whether ChatGPT, Claude, Perplexity, Gemini and Google name you, and get the plan to become the answer.

Check my AI visibility →
Sundays only

Get the Sunday newsletter.

One email a week. AI experiments, marketing tactics, and the workflows Lilach is building right now in her own business.

Subscribe free

Let’s get your marketing running on AI.

Book a free 30-minute call

We figure out what you need, where AI fits in, and what working together would look like.

Book the call →

Or take the 30-second calculator

You’ll see the hours and the money quietly leaking out of your week, and the three workflows worth building first.

Take the calculator →

Or grab the free AI resource library

Prompt packs, templates, checklists, and swipe files. The exact tools I build for paying clients. Yours, free.

Get the library →
Keep reading

More from the blog.