Why AI Often Recommends the Smaller Competitor

A solo consultant with 40 pages on her site keeps getting named by ChatGPT for queries her old agency, with 4,000 pages and a Forbes Council badge, never shows up for. She is not the exception. She is the pattern.

Key takeaways

• AI engines do not rank by brand size or backlink count. They rank by who gives the clearest, most specific first-party answer to the buyer's question.
• BakingSubs is a niche site with no ads, no paid backlinks, and no social presence. It has earned 162,500 Microsoft Copilot citations to date, with 112,500 of those landing in just the last three months.
• A bigger competitor's homepage usually covers ten topics shallowly. A smaller specialist's site covers one topic deeply. ChatGPT picks the second one almost every time.
• The "smaller competitor advantage" is a structural feature of how language models retrieve. It is not a temporary glitch that big brands will fix.
• The fix for an invisible site is not more content or more backlinks. It is narrower scope, clearer author signals, and direct answers to the exact questions buyers type into AI engines.

Why size stopped mattering when buyers moved to AI search

AI engines do not pick the biggest site. They pick the site that most clearly answers the question being asked. Those are not the same thing, and they used to be confused with each other because Google's old ranking system rewarded both at once.

When a buyer types "best business coach for second-time founders in Austin" into ChatGPT, the model is not browsing the web in real time and counting backlinks. It is pulling from sources where that exact buyer question (or something very close to it) has a clean, specific, first-party answer attached to a real human or brand. A big agency's homepage says "we help founders scale." A specialist's site has a page titled "Coaching for second-time founders: what is different the second time around" written by the named coach. Guess which one gets cited.

This is the part most people miss. The smaller competitor is not winning despite being smaller. The smaller competitor is winning because being smaller forced them to be specific. Specificity is the actual ranking signal now.

What ChatGPT, Claude, and Perplexity actually weigh

The four engines that matter for buyer research are ChatGPT, Claude, Perplexity, and Microsoft Copilot. They behave differently on the margins, but the core selection logic is the same.

They reward:

• A direct answer near the top of the page. Not a hero image and a "book a call" button. An actual sentence that answers the question the page title implies.
• Author identity that is real and easy to verify. A named person, a Person schema tag, a LinkedIn link, a photo. Claude weighs this especially heavily. A nameless "team" page loses to a single human with a face.
• Topical narrowness. A site that covers one thing in depth beats a site that covers ten things shallowly. Perplexity, in particular, surfaces specialists.
• Clean page structure. One H1 per page, H2s that read like the questions buyers ask, content that does not bury the lede.
• First-party experience. Pages where the author is clearly the practitioner, not a content marketer summarizing someone else's work.

Now look at what they do not weigh much:

• Total backlink count
• Domain age (past a basic threshold)
• Social media following
• Brand recognition outside the specific niche
• Total page count or word count of the site

That last bullet is where the smaller competitor advantage comes from. A 4,000-page site has 4,000 chances to dilute its signal. A 40-page site, if every page is specific and on-topic, has 40 chances to be cited as the clearest answer.

A real example: the solo consultant beating the big agency

Here is a scenario that plays out constantly. A reader emailed me about something close to this last month.

Imagine a fractional CFO named Renata in Pittsburgh. She works with manufacturing companies between $5M and $25M in revenue, mostly family-owned. Her site has 32 pages. The pages are titled things like "How a fractional CFO works with a third-generation manufacturing owner" and "What family-owned manufacturers should ask before hiring a fractional CFO." She wrote them. They have her name, her photo, her LinkedIn. Each post answers one specific question.

Her old firm, a 60-person consulting agency she left two years ago, has a site with hundreds of pages on every topic in finance and operations. They show up for nothing in ChatGPT when a manufacturing owner asks about fractional CFO help.

Why? The agency's pages are written by content marketers. They cover "10 things every CEO should know about cash flow." They are not signed by a real practitioner. They cover every industry, so they cover no industry. When ChatGPT is asked who helps third-generation manufacturers in the Rust Belt with cash flow, the agency's site does not surface as a specific answer. Renata's does.

This is not magic and it is not unfair. It is the engines doing exactly what they are designed to do: match a specific question to a specific answer from a credible source.

The BakingSubs proof: a niche site out-cited far larger sites

I built BakingSubs as the proof of concept for The Citation Cluster Method. It is a niche site about baking ingredient substitutions. No ads. No paid backlinks. No social media presence. No mailing list popups.

To date it has earned 162,500 Microsoft Copilot citations. 112,500 of those landed in just the last three months. That is not a slow-build curve. That is compounding. At peak it pulls over 5,000 daily Google clicks alongside the AI citations.

There are dozens of much larger baking sites, recipe portals with millions of monthly visitors, with celebrity authors and full editorial teams. They get cited too, sometimes. BakingSubs gets cited more for substitution queries because every page on BakingSubs answers one substitution question with one clear, named, first-party answer.

The lesson here is not "go start a baking site." The lesson is that the same structural choices work in any expert-led niche. The full breakdown of how the site got there covers the specific choices that compounded.

The contrarian thesis, stated plainly

Most marketing advice for solo experts says: build authority by getting backlinks from big sites, guest posting, getting on podcasts, and growing your following until the algorithms favor you.

That advice is now actively wrong for AI search. I will say it plainly: a solo expert with 30 specific pages on one narrow topic beats a 4,000-page brand site in AI citations almost every time, and chasing backlinks or follower counts is the wrong response to being invisible. The right response is to get narrower and clearer.

You do not need to out-resource the bigger competitor. You need to out-specific them. A bigger site cannot easily get narrower, because narrowing means killing pages, killing departments, killing content streams that exist for non-AI reasons. You can narrow tomorrow. That is the leverage.

Why bigger sites cannot easily catch up

The structural reason this advantage holds is that bigger brands have committed to breadth. Their site exists to serve many audiences, many products, many search terms. Their homepage is a compromise between every stakeholder. Their content calendar is set by SEO teams chasing volume.

To win in AI search, they would need to:

• Pick one buyer question per page and answer only that question
• Attach a single named human author to every expert-claim page
• Kill thin, generic, evergreen "what is X" pages that dilute their topical focus
• Restructure their site so one topic owns one clear path

A solo expert can do all four in a weekend. A 200-person company cannot do it in a year. That asymmetry is why smaller specialists keep showing up in ChatGPT recommendations while bigger competitors do not, even when the bigger competitor has every traditional advantage.

What to do this week if you are the smaller competitor

If you are reading this and you are the specialist, here is the order of operations.

1. Find out what ChatGPT actually says when a buyer asks for someone like you. Run the AI Visibility Check. It asks 8 discovery-intent questions per engine and tells you which of the four buckets you are in: Invisible, Mixed, Winning, or Empty-niche.
2. Pick the one buyer question you most want to be cited for. Write the page that answers it under your name. One topic. One human. One clear answer near the top.
3. Build three more pages around the same topic. Not ten. Three good ones. This is the start of a citation cluster. The full mechanic is in The Citation Cluster Method.
4. Check again in six weeks. You are looking for movement on a few specific queries, not a flood. First citations typically show up before traffic does.

Frequently asked questions

Why does AI recommend smaller businesses over bigger ones?

AI engines pick the clearest, most specific first-party answer to a buyer's question. A smaller specialist usually has narrower content, a real named author, and pages built around the exact questions buyers ask. A bigger brand's content is usually broader, written by content teams, and not attached to a single practitioner. The smaller site wins on specificity, which is what the engines actually reward.

Do I need backlinks to get cited by ChatGPT?

No. BakingSubs has earned 162,500 Microsoft Copilot citations without paid backlinks, ads, or social. Backlinks help Google ranking but are not the primary signal AI engines use when deciding who to recommend. Clear topic focus, named authors, and direct answers to buyer questions matter more.

My competitor has way more traffic than I do. Does that mean they will always win in AI search?

No. Traffic is a result of ranking, not a cause of citation. AI engines do not see your analytics. They see your pages. If your pages answer one buyer question clearly under your name and your competitor's pages cover ten topics shallowly, you can catch up faster than you think.

How do I know if I am actually invisible in AI search?

Open ChatGPT and ask the 5 to 8 questions a buyer would ask before hiring someone in your niche. If your name does not appear in any answer, you are invisible. The free AI Visibility Check runs this for you across the four major engines and tells you which bucket you are in.

Is this advantage going to disappear once big brands catch on?

Probably not soon. Bigger brands cannot easily narrow their content because their site exists to serve many audiences and product lines. A solo expert can restructure in a weekend. A 200-person company cannot. That structural gap is why the smaller-competitor advantage has been getting stronger, not weaker, as AI search grows.

Where to start

Stop comparing your traffic, your follower count, or your domain authority to the bigger competitor. Those are not what the engines are scoring. Open ChatGPT, ask three questions a buyer in your niche would ask, and see whose name shows up. If it is not yours, that is fixable, and the fix is not "be bigger." The fix is to be the clearest, most specific answer to one question that matters. Run the visibility check, pick that one question, and start there.