Type "best plumber in Chelmsford" into ChatGPT right now, before you read any further. If your firm has fifty five star reviews and a van booked solid three weeks out, you might reasonably expect to show up in the answer. A surprising number of business owners run this test for the first time and get a genuinely unpleasant surprise.
Ask the question you'd expect your customers to ask
Start embarrassingly simple. Open ChatGPT, Gemini, and Microsoft Copilot in separate browser tabs. Type the exact phrase a customer would type, not the phrase you'd use internally to describe your own business. "Removals company Harlow" produces different results from "who should I use to move house in Harlow." Try both versions. Then try one with real intent baked in — "trustworthy electrician near Basildon who actually answers the phone quickly." Wording matters more than most owners assume, because these tools are answering a specific question rather than matching a keyword string the way old search engines did.
Write down what comes back, word for word, in an actual document rather than relying on memory. You'll want that record later, once you've made changes, to see whether anything has genuinely shifted.
The free checks worth five minutes of your time
You don't need paid software for this first stage. A handful of free checks tell you most of what you need to know fairly quickly:
- Search your exact business name in each AI assistant and see whether it describes your services correctly
- Ask "who are the top five [your trade] companies in [your town]" and note whether you appear at all, and where
- Check whether the assistant cites a source — if it does, click through and read exactly what page it pulled from
- Repeat the same questions a week later, because the answers genuinely do shift over time
That last point tends to surprise people most. AI answers aren't static the way a Google Business Profile listing is. We've watched a Southend café disappear from one week's answer entirely and reappear the next, with absolutely nothing changing on their end at all — no new content, no new reviews, nothing.
Keep a running note of dates too. If you can see a mention appeared the same week you published a new FAQ page, that's useful evidence for what actually worked, rather than a guess you make six months later when you've forgotten what changed and when.
What a mention actually looks like once you find one
Being named isn't a simple yes or no. Sometimes an assistant mentions your business by name but gets a detail wrong — the wrong opening hours, a service you quietly dropped two years ago, a phone number that's been dead since you moved premises last spring. That's arguably worse than not appearing at all, because a customer trusts the answer and then gets frustrated with you specifically, not with the AI tool that misled them. Check carefully for accuracy, not just presence, every single time you run this test.
Other times a competitor gets named outright and you don't, despite your site having more content and visibly better reviews on paper. This usually comes down to how clearly a competitor's pages answer specific, narrow questions, which is something we build into our own search visibility work across Essex. It's rarely, in our experience, about who has the bigger or more expensive-looking website.
Why a website that looks "fine" doesn't help here at all
A lot of business owners point to a smart-looking website as proof they're already covered. It isn't, not for this particular problem. AI assistants are pulling structured, specific, easily quoted information — pricing ranges, exact service areas, direct answers to direct questions — rather than rewarding a nice homepage photo and a vague paragraph about "our commitment to excellence." We've reviewed sites that cost several thousand pounds to build and genuinely couldn't tell you, from the visible text alone, what town the business even serves, let alone what it actually charges.
A smart website and a visible website are two different things entirely, and plenty of Essex businesses have quietly paid a premium for the first without ever getting the second.
A quick example worth stealing: the gym that showed up for everything except itself
One Grays-based gym we spoke with had decent, well-established local SEO. Showed up fine on Google Maps, ranked reasonably on Google itself. But ask any AI assistant "gyms with personal training near Grays" and they weren't mentioned once across three different tools we tested, over several separate weeks. It turned out their site described classes and equipment in long, flowing paragraphs but never once stated, in a clean standalone sentence, "we offer one-to-one personal training sessions in Grays." The information technically existed somewhere on the page. It just wasn't written in a form an AI model could lift and quote with any confidence.
Checking across more than one assistant genuinely matters
Don't stop at ChatGPT simply because it's the one everyone talks about at dinner parties. Gemini draws heavily on Google's own index and often behaves quite differently. Copilot leans on Bing's results underneath it. Perplexity cites sources far more explicitly than most of the others, which actually makes it a particularly useful diagnostic tool — if Perplexity names a specific competitor page by title, go and read that exact page, because it's doing something right that yours probably isn't doing yet.
We test across at least three assistants for every client we take on, because a business that's invisible on one tool but perfectly fine on another needs a genuinely different fix than one that's invisible absolutely everywhere.
Keep a simple spreadsheet, one row per assistant, one column per question. It looks basic. It's also the clearest way we've found to spot patterns — a business invisible on Gemini specifically, for instance, often has a thin or outdated Google Business Profile sitting behind it, since Gemini leans so heavily on that particular data source.
What to actually do once you've found the gaps
If you've run these tests and come back with nothing at all, resist the urge to panic and rewrite the entire site in a single frantic afternoon. Pick the two or three questions customers ask most often — not what you assume they ask, what they actually type or say out loud — and make sure a page on your site answers each one directly, near the top, in plain, confident sentences. One question answered properly per section tends to work far better than cramming everything into a single sprawling page. For a lot of the smaller Essex businesses we've worked with, this kind of targeted restructuring delivers more than a full site redesign ever would, and it's a core part of how our visibility audits begin for every new client.
Don't try to fix everything at once, either. We've seen owners attempt a full content overhaul in one weekend, get exhausted by Sunday afternoon, and abandon the project halfway through with half the site rewritten and half left untouched. Three focused pages done properly beats twelve pages rushed, every time.
The Bottom Line
Testing whether AI assistants know your business exists takes less time than most owners expect, and it costs nothing beyond a few minutes of typing into a browser. What it reveals, though, often changes how a business thinks about its own website entirely — less a brochure to admire, more a set of clear, confidently quotable answers waiting to be found. Run the test this week. Write down exactly what you find. Then fix the gaps in the order customers actually ask those questions, not in the order that happens to feel most urgent from your side of the desk. It's a cheap test, an afternoon of actual work, and genuinely one of the clearer early-warning signs available to a small Essex business right now.
Looking for SEO agency in Essex?
Let's talk about what your business actually needs
Digital Impact Ventures provides SEO, AI visibility and digital marketing support for businesses across Colchester and Essex. Get in touch and we will talk through where your business is now and where it needs to go.