When a rival gets the recommendation

ChatGPT recommends a competitor for our main service: what do we do first?

Evangelos Pantidos · WhoAINames · 15 October 2026 · 9 min read

Why does ChatGPT recommend them and not us?

Because the sources it reads agree on them more clearly than on you. When a buyer asks “Who’s the best {service} firm in {city}?”, the assistant doesn’t judge firms on quality the way a client would. As far as anyone can tell from the outside, it looks at what it can find that matches the question (lists, directories, reviews, articles, firms’ own pages) and names the firms that come up most consistently.

So a rival can be named because they’re on the “best {service} firms in {city}” lists, their directory profiles say exactly what they do, or their own site answers the buyer’s question directly. None of that requires them to be better than you. It requires the web to be clearer about them than about you.

That’s also why it can feel random. A smaller firm that does one thing, and is described the same way everywhere, can beat a larger firm whose specialism is one line in a long services menu. AI rewards being clearly and repeatedly associated with the service, not being the biggest.

Which sources is it using?

Mostly lists, directories and review platforms, plus a few pages that answer the question directly. The best published evidence comes from outside professional services, so read it as a pattern, not a rule for your sector:

• Lists come first. Ahrefs studied 26,283 source URLs behind ChatGPT’s answers to 750 recommendation prompts (software, products and agencies). “Best X” blog lists were 43.8% of all the page types cited, the single largest group. The study also found that brands placed high on third-party lists were more likely to appear in the answers.
• Ranking platforms count as lists too. The same study counted sites such as G2 and Clutch separately, as lists that aren’t blog posts. For IT and consulting firms, they can play the role a directory plays for a law firm.
• LinkedIn matters more than most firms think. In a 2026 Semrush study of 325,000 prompts, run with LinkedIn, LinkedIn was the second most cited site on ChatGPT Search, Google AI Mode and Perplexity. On average 11% of AI responses referenced it. On ChatGPT Search and Google AI Mode, 59% of those LinkedIn sources were individual members rather than company pages, and about 95% were original posts, not reshares.

You can check your own case directly. Perplexity and Google AI Mode show their sources under each answer, and ChatGPT shows them when it searches the web. Ask the question your buyer asks, then click every source behind the rival’s name. That list is your to-do list.

What it looks like: one firm’s source map

Here’s a real check on a national UK law firm, anonymised. We asked 20 buyer questions about its services, each once on four AI engines. The firm was named in 13 of the 80 answers, and the report identified one rival firm as its top competitor. These are the sources the engines cited instead of the firm’s own site:

The third row is the one firms miss. A rival firm’s website was among the sources AI used to answer buyers’ questions, and the firm we checked, which is far better known, wasn’t. A competitor that answers the buyer’s question on its own site becomes part of the answer, even when the question isn’t about them.

This check is legal, the sector we’ve run in full so far. The shape is the same in other sectors with different names in the first two rows: accountancy directories and award lists, review platforms such as Clutch or G2 for IT and consulting firms, and sector “best of” lists everywhere.

We rank well on Google. Doesn’t that count?

It helps, but it isn’t the same result, so measure both. Google ranks pages. An AI assistant recommends firms, and it builds that recommendation from many pages at once, many of them not yours. A firm can sit on page one for “{service} {city}” and still be missing from the AI answer to the same question, because the lists and directories behind the answer don’t mention it.

Page one is also worth less than it was, because Google now puts its own AI answer above it. Ahrefs compared 300,000 keywords before and after Google’s AI Overviews arrived: when an AI Overview appears, the top-ranked page gets 58% fewer clicks on average, and the page in position two about half as many (50.8% fewer). Pew Research tracked what 900 US adults actually did on Google. When an AI summary appeared, they clicked a normal result in 8% of visits, against 15% when there was none, and clicked a link inside the summary in just 1%. Both are US studies of Google, and Ahrefs measured informational searches, so the effect on “best {service} in {city}” will differ. Page one still matters, but more and more buyers read the answer above it and stop there. Who that answer names is a separate result.

The useful test is simple: for each of your main questions, write down your Google position and how many times AI names you across a few runs. Where the two disagree, the gap is outside your website.

What should we fix first?

Work through the sources in this order, one service at a time.

1. Be listed where the answer comes from. Take the sources behind the rival’s name and check whether you’re on each one, for this service. For a law firm that’s often a directory ranking. For an accountancy it may be an award shortlist or a “best accountants for {type of client}” article. For an IT firm or consultancy, a review platform profile. Being missing from a list AI keeps citing is often the clearest gap to start with.
2. Be described the same way everywhere. Same service name, same specialism, same city, same type of client, on your website, directory profiles, review platforms, partners’ LinkedIn profiles and your Google Business Profile. If one place says “commercial disputes” and another says “litigation”, AI has two weak signals instead of one strong one.
3. Answer the buyer’s question on your own site. One page per service that answers the question in its first sentence, uses the questions clients actually ask as headings, and backs its claims with something only you have: outcomes, fee ranges, timelines, credentials, sectors served.
4. Get others to say it. Reviews that mention the service by name, a partner quoted in the trade press, original LinkedIn posts from the partners who lead the service (not reshares), and a place on the lists AI already reads.

Don’t start by rewriting your homepage. Start with the sources the answer already uses. Then move to the next service, because AI judges each one separately.

How long before it changes?

Months rather than weeks for most questions worth winning, and nobody can promise a date. It depends on how contested the question is. A narrow question, such as a specialist service for a specific type of client, can move sooner. “Best {service} in {city}” in a big city, where several established firms compete, takes longer.

Two things make it slower than people expect. Directories and lists update on their own schedules, and AI assistants don’t all read the same sources, so a change can show up on one before the others. Check the same questions every month, across more than one assistant, and judge the trend, not one answer.

Should we ask our SEO agency?

Yes, but ask specific questions, because this overlaps with SEO without being the same job. A good SEO agency already works on your site, your Google Business Profile and some listings. What’s often missing is measuring the AI answer itself and working on the outside sources behind it.

Four questions to put to them:

• Do you track how often AI assistants name us for our main questions, by service and city?
• Which sources do those answers cite, and are we on them?
• Who is named instead of us, and why?
• What will you change in the next 90 days, and how will we see whether it worked?

If the answers are vague, the work isn’t being done yet, whoever ends up doing it.

Questions partners ask

Sources

1. Ahrefs (Glen Allsopp), “Do Self-Promotional ‘Best’ Lists Boost ChatGPT Visibility? Study of 26,283 Source URLs”, 4 December 2025, 750 prompts about software, products and agency recommendations. ahrefs.com (not a study of professional service firms).
2. Semrush, LinkedIn AI visibility study, 10 March 2026, 325,000 prompts on ChatGPT Search, Google AI Mode and Perplexity, January–February 2026, run in collaboration with LinkedIn. semrush.com
3. Ahrefs (Ryan Law), “AI Overviews Reduce Clicks by 58%”, 4 February 2026, 300,000 keywords, Google Search Console desktop click-through rates, December 2023 vs December 2025. ahrefs.com (informational keywords; a correlation, not proof of cause).
4. Pew Research Center, “Google users are less likely to click on links when an AI summary appears in the results”, 22 July 2025, browsing data from 900 US adults, March 2025. pewresearch.org