Our business does not come up when I ask ChatGPT who the best brokers are

Why does ChatGPT not know about us?

Because AI answer engines build their answers from sources they can read and, crucially, sources that are not you.

When somebody asks for a recommendation, the engine is not searching your website and deciding you seem decent. It is assembling an answer from directories, comparison pages, trade press, review platforms and, yes, company sites. Your own pages count, but they count as you describing yourself, which carries less weight than somebody else describing you.

We measured this on our own business recently. Across 408 URLs the engines cited in one test, ten were ours and not one of the other 398 was a third-party page about us. Plenty of self-description, no outside corroboration. That is the shape of the problem for most firms, and it is almost certainly the shape of yours.

Is it because our website is bad?

Usually not. Most broker websites look fine and read fine to a human.

The question is whether a machine can extract a clear answer from them. Three things decide that, and all three are invisible to you as a visitor.

Whether your content is in the page when it arrives, or gets assembled by JavaScript afterwards. Many crawlers do not run scripts, so a site that builds itself in the browser is a blank page to them.

Whether your headings are questions people actually ask, or labels like "Our Services" and "Why Choose Us". Engines extract from question-shaped headings because that is what maps onto what was asked.

Whether you have structured data telling a machine what your business is, where it operates and what it sells. Without it, the engine is guessing from prose.

None of that is about how the site looks. All of it is about whether it can be read.

What actually gets a broker named?

Specificity, in our experience, more reliably than size.

Here is a pattern worth knowing. In a recent measured test across four engines, we were named on the sector-specific and product-specific questions and absent from every generic one. Named on motor claims and insurance brokers. Absent on "best AI consultancy for UK SMEs".

That is not a coincidence and it is not unique to us. The generic questions are dominated by directories and roundup articles, and no page you publish on your own site will beat those. The specific ones are winnable, because they reward being genuinely the right answer to a narrow question rather than a plausible answer to a broad one.

For a brokerage that means you are far more likely to be named for "commercial insurance broker for haulage in the North West" than for "best insurance broker UK". Which is fine, because the first one is a better lead anyway.

What are the three things to check first?

Ask the engines yourself, properly. Not "who are the best brokers", which is too broad to be useful. Ask the question a real client would ask, in their words, including the sector and the location. Do it in a fresh session with no account signed in, because otherwise you are seeing answers shaped by your own browsing history. Write down what comes back and who gets named.

Check whether your site can be read without JavaScript. If you are not technical, ask whoever built it a direct question: does the full page content appear in the HTML before any scripts run? A yes or no will do.

Look at where you are mentioned that is not your own website. Directories, trade press, review platforms, LinkedIn. If the honest answer is nowhere, that is your actual problem, and it is a slower fix than the technical one but it is the one that moves the needle.

When is this not worth fixing?

If you get most of your business through introducers, referrals and existing relationships, and you are not trying to grow beyond that, then this matters much less than somebody selling it to you will suggest.

AI visibility is worth investing in when a meaningful share of your new business could plausibly start with somebody searching who does not already know you. If that is not your model, spend the money elsewhere. We have told clients exactly that.

What we do about it

We run a measured test of the queries your buyers actually use, across ChatGPT, Claude, Google AI Overviews and Perplexity, and show you who got named instead of you. Then we tell you which gaps are technical, which are content, and which need outside mentions you will have to earn.

The AI Visibility Audit at the AI Visibility Audit page is £995 and is delivered in five working days. Most of them surface two or three things you can fix yourself in an afternoon without hiring anybody, and those go at the top of the report rather than behind a proposal.

If you want the background first, our AI visibility audit guide explains the method, and our guide on how AI decides who to recommend explains the mechanics. For the sector view, our insurance brokers page covers where else this fits in a brokerage.

Frequently Asked Questions

Your next step

Take the free two-minute Quick Check for an AI readiness score, see what the £995 AI Visibility Audit covers, or book a free 45-minute discovery call. Definitions for the terms used here are in the AI glossary and the Claude glossary.