Fundamentals · 5 September 2026

By Eddie Eastwood, Founder

Why Does ChatGPT Recommend My Competitor and Not Me?

Why does ChatGPT name them and not you?

You and a competitor sell close to the same thing, to the same buyers, in the same area. You ask ChatGPT a question a real customer would ask, and it names them. Not you. This is a sharper problem than not being known at all: the engine clearly knows who operates in your category, and it named someone else.

In a single-query measurement run for this page, five identical ChatGPT runs on 2026-09-04 asked "Which company should I call for emergency boiler repair in Leeds?" Between them they named 13 distinct businesses, and seven of the 13 (53.8%) did not appear anywhere in Google's organic top 20 for the same search. Three of the four businesses named in every one of the five runs were absent from the top 20 entirely; the fourth ranked 20th.^[1] Ranking on Google and being the business an engine names are not the same event; Ranking No.1 on Google Won't Get You Into ChatGPT covers how retrieval and citation separate. Most of the field already carries some of what an engine looks for: a separate first-party study found 92.2% of 179 successfully audited US local service business websites were rated at least partially ready to be recommended by AI systems (10-signal audit, successful fetches only, N=179).^[3] Why Does AI Name Businesses That Seem Worse Than Yours? covers why documentation rather than quality decides which business gets named. This page starts one step further along, from the four things that differ between you and the competitor being named.

The four things that decide it

Four factors explain why one of two comparable businesses gets named and the other doesn't.

A citation cluster that already agrees. Directories, review platforms, comparison articles and local write-ups cite the same handful of businesses together when they cover a category. AI Recommends Businesses in Clusters covers the mechanism in full. In short: if your competitor already appears alongside two or three other sources that name them together, an engine reading that cluster has independent-looking confirmation. If you've never entered that cluster, there's nothing for the engine to cross-check you against, even if your own website is excellent.

Facts that read the same everywhere. Name, address, phone number, service list, hours, all matching across your own site, your business listings, and anywhere else you're mentioned. A competitor whose details are identical everywhere gives the engine one clean, uncontradicted fact set to draw on. A business whose old address still shows up on one directory, or whose phone number differs between two listings, gives the engine a reason to hedge, and hedging means naming the business with no contradictions instead.

Reviews that are recent, not just plentiful. A large batch of reviews that stopped two years ago reads as stale. A 2026 survey of 1,002 US consumers found that a review posted within the last month outweighs a high star rating alone for trust (44% versus 42%), and that consistent sentiment across multiple reviews matters most of all (56%).^[2] A smaller, steadily arriving set of recent, consistent reviews reads as current and active in a way a large batch that stopped growing does not.

A page that answers the buyer's opening question directly. If a real buyer's first question is "who does X in Y", and your competitor has a page built to answer exactly that question while your equivalent page is a generic services list, the competitor's page is easier for the engine to lift a direct answer from. The lever is one page that matches one question, in the buyer's own phrasing, rather than more content overall.

How to find out which of the four is costing you

Before fixing anything, find out which factor is in play. Guessing wastes effort on the wrong one.

Run the comparison prompt yourself. Ask the exact question a buyer would ask, phrased the way they'd phrase it, on the platform they'd use. How to Check If AI Recommends Your Business covers building a proper prompt list and reading the results; for this diagnostic, one prompt run several times is enough to see whether the competitor is named consistently or only sometimes.

Read what the answer cites, not just what it says. If the platform shows sources or you can trace the likely ones (directory listings, review platforms, a comparison article), check whether your competitor appears across several of them together. That's the citation-cluster signal. In the same single-query measurement (the Leeds boiler-repair search run for this page on 2026-09-04), Perplexity retrieved 16 sources per run and named only 2 or 3 businesses from them: 14 of 16 retrieved sources (87.5%) went unnamed in one run, 13 of 16 (81.2%) in the other.^[1] Being retrieved puts a business in the candidate pool; something else decides who gets named from it, which is exactly what these four factors are about.

Pull your own facts and theirs side by side. Your website, your Google Business listing, and any directory that lists both of you. Look for the small mismatches: an old suite number, a phone number one digit off from a rebrand, a service you added that one listing never picked up.

Check review dates, not just review counts. If your last review is older than theirs, or your volume has flattened while theirs is still climbing, that's a recency gap even if your total is higher.

Search your own site for the buyer's exact question, then search theirs. If their page carries that question close to the wording a buyer would use and yours only touches the subject somewhere on a services list, that's the opening-question gap.

Fix order

Cost and dependency set the order below. Each step is cheaper than the one after it, and each clears a blocker the next one relies on.

1. Fix fact contradictions first. It costs time, not money, and every other fix depends on the engine already having a clean, uncontradicted picture of who you are and what you offer.
2. Build the opening-question page. Write the page that answers the buyer's actual first question, in close to their own words, once your facts are consistent enough to back it up.
3. Address review recency. A steady process for asking recent customers for a review does more for this factor than chasing a higher total count.
4. Work into the citation cluster. This is the slowest of the four, because entering a cluster waits on other sites deciding to name you. How to Get Your Business Recommended by ChatGPT and AEO for Local Businesses cover the practical steps for building that presence.

Doing them in this order means each fix removes a real reason the engine had to prefer the other business, rather than adding effort somewhere the engine was never actually looking.

What this doesn't fix

None of this guarantees a swap. Closing those four gaps removes reasons the engine currently has to prefer your competitor. It doesn't buy a placement on any single run. A generated answer can vary between two identical requests, however strong your underlying signals are. In the same single-query measurement run for this page, only 4 of the 13 businesses named appeared in all five identical runs. Another 4 of the 13 appeared in exactly one.^[1] If your business is newer, in a smaller market, or genuinely doesn't have the track record your competitor does yet, this diagnostic tells you what to build toward, not a shortcut past it. And if the real question is closer to "why doesn't my business come up at all", rather than "why do they beat me specifically", that is the general invisibility diagnostic: Why Doesn't My Business Come Up When People Ask AI for Recommendations? works through it in check order.

FAQ

Why does ChatGPT recommend my competitor and not me?

It comes down to one of four things: the competitor sits inside a cluster of sources that already agree on who the option is; their business facts read the same everywhere the engine looks; their reviews are recent as well as numerous; or they have a page that directly answers the buyer's first question and you don't. The engine is deciding what it can confirm, and confirmation is built from consistent information across independent sources.

What is a citation cluster, and why does it matter for AI recommendations?

A citation cluster is the group of sources, directories, review sites, comparison pages and articles, that already name the same business together when they discuss your category. An engine drawing an answer from that cluster keeps naming whoever the cluster already agrees on, because agreement across independent sources reads as a safer answer than a single unconfirmed one.

Does having more reviews than my competitor guarantee I'll be recommended instead?

No. Volume alone doesn't decide it. BrightLocal's 2026 survey of 1,002 US adults found recency and consistency count for more than raw star rating: a review posted within the last month outweighs a high star rating alone (44% versus 42%), and consistent sentiment across multiple reviews matters most of all (56%) [2]. If your review count is higher but your last review is old, a competitor with fewer, fresher and more consistent reviews can still look like the safer, more current answer.

What's the fastest thing I can fix if a competitor keeps beating me in AI answers?

Fix contradictions in your own business facts first: name, address, phone number, service list and hours, checked across your own site, your listings and any directory that mentions you. It costs nothing but time, it removes a reason for the engine to hedge, and it's the precondition every other fix depends on.

Will fixing this guarantee ChatGPT recommends me instead of my competitor?

No, and no page on this topic can promise that. Closing those four gaps removes the reasons an engine currently has to prefer the other business. It doesn't buy a guaranteed placement, because no one controls what a generated answer names on any single run.

Where do I go next to build AI visibility from scratch?

This page assumes you're already a real option and diagnoses why a specific competitor is being named instead. For the full step-by-step process of earning a first AI recommendation, see the practical guide at How to Get Your Business Recommended by ChatGPT.

Short version, restated

Run the diagnostic against one named competitor, not against the market. These four factors are comparative: you are looking for the single place their record is easier to confirm than yours. Close that one, then re-run the same prompt before touching anything else.

Your competitor's advantage on any of these four is checkable, one at a time: Check whether AI platforms can find and cite your business →