What actually happens in the half-second between someone typing “best project management software for small teams” into ChatGPT and the answer appearing? Not a search of a quality database, because none exists. Not a human judgment, because there is no human. The engine reconstructs a consensus. It scans what it has absorbed from across the web, looks for the names that credible sources repeatedly call the best, and hands back the ones with the strongest agreement behind them. Understanding that process is the whole of appearing in AI search for ‘best of’ queries, and almost nobody outside a small group of practitioners understands it yet.
This matters because ‘best of’ queries are where buying decisions get made, and they are moving from Google’s ten blue links to AI answers that name two or three options and stop. If your brand is one of the named few, you capture demand at the moment of decision. If it is not, you are invisible at the exact instant a customer is choosing. The mechanics of how AI search handles ‘best of’ queries are learnable, and they reward a specific kind of work that most brands are not doing.
The AI is not judging, it is synthesizing
The first thing to internalize is that an AI engine has no independent opinion about what is best. It cannot try your software, taste your product, or evaluate your service. When it answers a ‘best of’ query, it is reporting a synthesis of what other people have already said. The engine is a mirror held up to the web’s collective judgment, and the names it returns are the names that judgment most agrees on.
I ran the test to watch it work. Asking ChatGPT for “the best email marketing tools for creators,” the answer named several tools and, crucially, justified each with synthesized reasoning: this one is known for automation, that one for ease of use, another for its pricing. Pushing further, asking why it chose those specific tools, the answer was revealing. It explained that these were the tools most frequently recommended across reviews, comparison articles, and user discussions for that use case. It was not claiming to know they were best. It was reporting that the web’s consensus pointed to them.
That distinction changes everything about how you approach AI search and ‘best of’ queries. You are not trying to convince a machine that you are excellent. You are trying to build the web-wide consensus that the machine reads and repeats. The engine is downstream of the conversation about your category. Win the conversation across enough credible sources, and the engine has no choice but to name you, because naming you is just accurately reporting what its sources say.
Consensus is the signal the engine trusts
Here is the concept to name, because it organizes everything else: consensus is the signal. AI search resolves a ‘best of’ query by finding the brands that many trusted, independent sources agree belong in the answer. One source calling you the best is an outlier the engine discounts. Twenty credible sources independently naming you a leader is a pattern the engine treats as truth. The work of AEO for ‘best of’ queries is manufacturing that pattern honestly.
This is why a single great page on your own site does almost nothing for these queries. The engine knows your site is biased toward you, so it weights your self-description lightly. What moves the needle is third-party agreement: review platforms, comparison articles, industry roundups, expert recommendations, community discussions, all independently pointing to you. Each one is a vote, and the engine counts votes across sources it trusts. When the votes concentrate on you, you become part of the consensus set, and the consensus set is what gets recommended.
The practical implication is that AI visibility for ‘best of’ queries is earned off your own property, in the wider ecosystem of who-recommends-whom. A brand obsessed with its own website while ignoring its presence in third-party roundups and reviews is optimizing the one surface the engine trusts least. Meanwhile a competitor quietly getting included in every credible ‘best of’ list for the category is building the exact consensus the AI reads. Between two brands of equal real quality, the one with more independent mentions wins the AI answer, because the engine cannot see quality, only agreement.
Why ‘best of’ listicles became critical
The single most direct input into how AI search handles ‘best of’ queries is the third-party ‘best of’ listicle. When a reader asks for the best in a category, the engine reaches heavily for exactly the human-written roundups that already answer that question, “10 best CRMs for startups,” “top 7 accounting tools for freelancers,” and synthesizes across them. Those articles are pre-built consensus, and the engine loves pre-built consensus because it is doing the same job the engine is doing.
This means being included in the credible listicles for your category is one of the highest-leverage moves in AEO. Every reputable roundup that names you is a strong vote in the consensus the AI assembles, and appearing across many of them makes your inclusion in the AI’s answer close to inevitable. The brands winning ‘best of’ queries in AI search are frequently the same brands that show up in the most human-written ‘best of’ articles, because the AI is reading those articles and repeating their verdict.
Getting into those roundups is its own discipline: identifying the articles that rank and get cited, reaching the writers and publications behind them, and making a genuine case for inclusion backed by what makes you distinct. It overlaps heavily with traditional digital PR, because the goal is the same, earning credible third-party endorsement, just aimed now at feeding an engine rather than only a human reader. A brand that treats ‘best of’ inclusion as a deliberate program rather than a happy accident builds AI visibility that compounds, because each new listicle strengthens the consensus and the consensus is what the engine speaks.
Make your content quotable, not just rankable
Consensus gets you into the set of names the engine considers. Quotable content is what the engine uses to describe you once you are in it. When AI search answers a ‘best of’ query, it does not just list names, it gives a reason for each, and those reasons come from the clearest, most specific descriptions of each brand available across the web, including your own content when it is structured to be extracted.
This is where your own site earns back some influence. Content that states plainly what you are best at, for whom, and why, in clean, specific, quotable language, gives the engine ready-made material to justify recommending you. Vague positioning, “we deliver innovative solutions for modern businesses,” gives the engine nothing to quote, so even if you make the consensus set, you get named without a compelling reason attached. A specific, honest claim, “built for solo bookkeepers who manage under 50 clients,” gives the engine a precise reason that maps directly to the ‘best of’ queries where you should win.
The two forces work together. Consensus, built through third-party mentions and listicle inclusion, gets you into the answer. Quotable, specific content, on your site and in the sources that describe you, shapes how favorably and precisely you appear once you are there. A brand that builds both, wide independent agreement plus clear extractable descriptions, dominates ‘best of’ queries in AI search, because it has given the engine both the votes to include it and the words to praise it. The engines are not gatekeepers you sweet-talk. They are reporters of consensus, and the brands that win are the ones that did the unglamorous work of becoming the answer everyone else already gives.