I asked five AI tools what they know about my business. Here's what I found.

I run a company that helps small businesses show up in AI search. Which means if I can't show up in AI search, I have a credibility problem.

So I decided to check.

I opened five AI tools and asked each one the same question: "What can you tell me about Clearmark Digital in St. George, Utah?" Then I documented exactly what they said — the good, the bad, and the blank stares.

Here's what I found.

The tools

I tested ChatGPT (GPT-4o, no web browsing enabled), Perplexity, Google Gemini, Claude (Anthropic), and Microsoft Copilot (Bing-powered). I ran every test on the same afternoon. Where tools had the option to browse the web, I left it enabled.

What ChatGPT said

Nothing useful.

ChatGPT told me it didn't have reliable information about Clearmark Digital and recommended I check the company's website directly. It wasn't wrong — it just didn't know anything. ChatGPT's training data has a cutoff, and Clearmark Digital is a new company. If we didn't exist in its training data, we don't exist to ChatGPT.

This is the invisible problem. ChatGPT has over 900 million weekly users. A meaningful percentage of them ask it questions like "who does AI consulting in Utah?" or "is there a good AI marketing agency near me?" If ChatGPT doesn't know you exist, you're not in those conversations.

What Perplexity said

Better — because Perplexity searches the live web.

It found our website and gave a reasonable summary: a St. George, Utah agency focused on AI tools and AI search visibility for small businesses. It cited clearmarkdigital.com as its source, which meant I could see exactly where the information was coming from.

It got the core right. But it described our services in generic terms — nothing that would distinguish us from a dozen other agencies. That's not a hallucination. That's a content problem. Perplexity can only summarize what it finds. If what it finds is thin, the summary is thin.

What Google Gemini said

Similar to Perplexity, with one difference: it also pulled from our Google Business Profile.

The GBP listing is complete and current, so Gemini had a better picture of what we actually do. It mentioned specific services by name. It correctly identified St. George as our location. The summary felt like it was describing the actual business rather than a vague approximation of it.

The lesson from Gemini: your Google Business Profile isn't just for people searching Google Maps. It feeds directly into Google's AI tools. A complete, well-maintained GBP is the fastest win you can get for AI visibility — and it costs nothing to update.

What Claude said

Also nothing useful, for the same reason as ChatGPT. Training data cutoff. We're too new.

What's interesting is that Claude was transparent about it — something along the lines of: "I don't have information about this specific company and wouldn't want to guess at details that could be inaccurate." I respect that more than a confident wrong answer. But the outcome for a potential customer asking about us is the same either way: no information.

What Microsoft Copilot said

The most detailed response of the five.

Copilot is Bing-powered, which means it searches the web and tends to index more broadly. It found our website, our Google Business Profile, and at least one third-party directory listing. It returned a multi-paragraph response that covered our services, our location, and our founder. One service description was paraphrased in a way that wasn't quite accurate — but overall it was the most complete picture any of the five tools produced.

What I actually learned

Running this test taught me three things I didn't fully expect going in.

Recency matters more than I realized. Two of the five tools knew nothing about us because we didn't exist in their training data. That gap closes over time as models are retrained — but for any business that launched in the last year or two, this is your reality right now. You may be invisible to the tools that don't search the live web, and there's no shortcut to fix that. You just have to be patient and make sure everything else is in order for when the models catch up.

Content depth is the gap live-search tools expose. Perplexity found us. What it found just wasn't very specific. The tools can only work with what's on your website, so if your website describes what you do in vague marketing language, that's what the AI will summarize. Plain-language, specific descriptions of your services — what you actually do, for who, and what the outcome is — give AI more to work with.

The Google Business Profile punches above its weight. Gemini cited our GBP services listing by name, which I wasn't expecting. Most business owners think of GBP as a Maps thing. It's not. It's one of the primary data sources Google's AI reads when it describes your business. Keep it complete, keep it current, and treat it like a second homepage.

What I changed after the test

We added more specific service descriptions to the website — not marketing copy, but plain-language explanations of exactly what each service does and who it's for. We also added a FAQ section with the questions clients actually ask before hiring us.

Perplexity's summary of our services improved within a week.

The thing most business owners don't do

They never run this test on their own business.

It takes 20 minutes. You don't need any tools or expertise. Open ChatGPT, Perplexity, Gemini, Claude, and Bing Copilot, ask each one what it knows about you, and write down what comes back. The gaps are usually obvious once you see them.

If your business is established and you're still coming up blank in most tools, that's a signal worth taking seriously. If you're new like us, some of this is just time — but there's still plenty you can do to accelerate it.