How to check if ChatGPT is recommending your competitors instead of you

Your analytics dashboard shows organic traffic, conversions and rankings. It shows nothing about the conversation happening when a buyer asks ChatGPT "who's the best supplier for X" and your competitor gets named while you don't. That conversation is invisible to every tool marketing teams have relied on for the last decade, which means most businesses have no idea whether they're winning or losing it.

This isn't a hypothetical concern for niche cases. Anyone researching a supplier, from an office manager comparing IT support providers to a procurement lead shortlisting logistics partners, is increasingly starting that research with a question typed into an AI assistant rather than a list of search results to click through one by one. If your business isn't part of the answer they get back, you've been quietly removed from consideration before your sales team ever gets a chance to make the case themselves.

Why you can't just check this in Google Analytics

When someone asks an AI assistant a question and gets a direct answer, there's often no click, no referral, no session. The interaction that decided your buyer's shortlist happened entirely outside anything your analytics platform can see. You can watch your organic traffic hold steady while your actual position in your market's buying conversations quietly erodes, and nothing in a standard dashboard will tell you that's happening until the enquiries themselves start drying up.

A manual process you can run today

You don't need specialist tools to get a first read on this. It takes about half an hour:

1. 01List the ten questions your buyers most likely ask before choosing a supplier in your category, phrased naturally rather than as keywords
2. 02Ask each one to ChatGPT, Google's AI Overviews, Copilot and Perplexity separately, using a fresh or logged-out session where possible to reduce personalisation bias
3. 03Record which businesses get named in each answer, in what order, and whether your business appears at all
4. 04Note which sources each tool cites, if it shows them, since that tells you exactly where the model is pulling its information from
5. 05Repeat the same set of questions monthly, since these answers shift as models are updated and as competitors do their own visibility work

This single exercise usually produces one of two reactions: relief that you're already showing up, or a fairly uncomfortable realisation that a named competitor has been quietly winning this conversation for months while you had no way of knowing.

Keep a simple record as you go: the question asked, the tool used, the date, and exactly which businesses were named in what order. Without that record, it's easy to remember the answers that felt notable and forget the pattern across all ten questions. A written log, even a basic spreadsheet, is what turns this from an anecdote you mention in a meeting into evidence you can actually track changes against month to month.

What to do with what you find

If you're not appearing at all, the fix isn't to write more content hoping something sticks. Work backwards from what the model actually cited for your competitors:

• If it cited a trade publication or directory, get your business covered there too
• If it cited a specific page on a competitor's site, look at what makes that page easy for a model to lift and quote, and apply the same structure to your own content
• If nothing gets cited and the model is answering from general knowledge, check whether your robots.txt is blocking AI crawlers like GPTBot, since that alone can remove you from consideration entirely
• If your brand is mentioned inconsistently across the sources that do exist, that inconsistency is likely diluting how confidently a model recommends you

Why this needs tracking, not a one-off check

AI answers aren't static. Models get updated, retrained and re-indexed, and a business that isn't mentioned this month might be next month if a competitor secures new coverage or fixes their technical access. Equally, a business currently winning this conversation can lose ground quietly if they stop investing while a competitor starts. Treating this as a one-off audit rather than an ongoing check means you'll only find out you've lost ground once it's already cost you enquiries.

The manual process above is a genuinely useful starting point, but it doesn't scale well past a handful of questions run occasionally. Dedicated AI visibility tracking exists specifically to run this at the scale and consistency that catching a shift early actually requires.

What good AI visibility actually looks like

It's not about appearing in every single answer. It's about being consistently in the conversation for the questions that matter most to your business, with accurate information attached when you do appear. A business showing up occasionally with wrong or outdated information can be worse off than one not appearing at all, since the model is actively spreading something incorrect about them.

Common objections, answered

We already track our Google rankings, isn't that basically the same thing?

No. Ranking well in classic Google results and being cited in an AI-generated answer are separate outcomes that don't automatically move together. A business can hold strong organic rankings while being entirely absent from AI answers, particularly if its content isn't structured in a way models can easily lift and quote.

Won't this just fix itself as AI tools get better at giving fair answers?

There's no evidence pointing that way. Models cite what they can access and trust, and businesses actively building citations and consistent brand information are the ones being named, not a random cross-section of every business in a category. Waiting for the models to become fairer on their own means waiting indefinitely while competitors who are actively working on it pull ahead.

Is it worth doing this if we're a smaller player in our market?

Often more so. Smaller, more specific businesses can be easier for a model to describe clearly and cite confidently than a large, broad competitor with a vaguer positioning. Being the clear, well-documented answer to a specific question is a genuinely achievable position for a smaller business, in a way that outranking a market leader for a broad Google search rarely is.

Building this into a regular habit

The businesses that get real value from this don't treat it as a one-off exercise before a board meeting. They run the same set of questions on a fixed schedule, keep a simple log of who gets named and cited over time, and treat a competitor's rising visibility as a signal to investigate rather than a coincidence. That habit, more than any single tool, is what turns this from an interesting one-off finding into something that actually protects pipeline.