The Well-Read Colleague Test: How AI Search Decides Who to Recommend

How AI search decides who to recommend — and what it ignores

A reasonable assumption is that AI search recommends the firm with the best website.

If your site is clear, modern, well-written, and technically sound, it should be the obvious source for ChatGPT, Perplexity, Google’s AI results, or any other answer engine when someone asks, “Who should I hire for this?”

Understandable. But incomplete.

AI search does not simply reward the best website. It is more likely to surface the firm that appears most consistently in the right context across the sources it can read.

That is the idea behind The Well-Read Colleague Test:

> If a well-read colleague in your market were asked to describe your firm, what would they confidently say — and does the internet contain enough evidence for AI to say the same thing?

AI search is not magic. It is pattern recognition over available evidence. It looks for repeated, credible associations between a company, a category, a problem, an audience, and a point of view.

1. Your website is a claim, not the whole reputation

Your website tells the market what you want to be known for.

That matters. But in AI search, your site is only one source of evidence.

An MSP might say:

> “We provide secure IT support for law firms.”

Useful. But if the rest of the public internet describes the company only as “IT support,” “managed services,” or “technology consulting,” the law firm specialization may not be strong enough to register.

A well-read colleague would repeat that positioning if they had seen the pattern elsewhere:

- case studies with law firms - articles about legal-sector IT risks - directory profiles mentioning law firm support - partner pages or event pages in the legal market - LinkedIn posts about law firm operations, compliance, or security

AI search behaves similarly. It does not need one perfect page. It needs corroboration.

Decision point: Before rewriting your homepage again, ask whether the claim on the homepage is supported anywhere else.

If not, the issue is not just web copy. It is weak market evidence.

2. AI looks for repeated context

A common reaction to AI search is, “How do we optimize for AI?”

The better question is, “Where does our expertise appear in context?”

AI search is less interested in isolated keyword placement than in consistent associations.

A consultant might publish one strong article called:

> “How to improve client onboarding.”

Useful, but generic.

Now compare that with a body of presence that repeatedly connects onboarding to a specific buyer situation:

- “Client onboarding problems in boutique consulting firms” - “Why agencies lose margin in the first 30 days of a new account” - “A checklist for handoffs between sales and delivery” - a podcast appearance on reducing onboarding friction - LinkedIn posts about onboarding scope creep after the sale

The second version gives AI search a clearer pattern.

The firm is not merely associated with “onboarding.” It is associated with onboarding inside a specific buyer environment.

Decision point: Do not ask only, “What topic should we cover?” Ask, “What topic should we be repeatedly associated with, for which buyer, in which situation?”

3. Authority includes where your signal appears

Traditional SEO trained businesses to think in pages, rankings, and keywords.

Those still matter. But AI search adds a broader question:

> Who else, besides you, gives your expertise context?

A B2B agency may publish strong thinking on category strategy. But if that thinking only exists on its own blog, the signal is thinner than it could be.

The signal becomes stronger when the same expertise appears across credible surfaces:

- client case studies - podcast interviews - conference speaker pages - trade publication quotes - partner ecosystem pages - review platforms - industry newsletters - LinkedIn posts engaged with by relevant peers

This is not about gaming the algorithm. It is about making your market reputation legible.

Decision point: If you want to be recommended for a specific problem, identify three credible places outside your website where that association should exist.

The question is not, “Can we trick AI into mentioning us?”

The question is, “Are we visible where a recommendation engine would reasonably look for proof?”

4. Compound Presence beats campaign bursts

AI visibility is not created by one busy week.

Many small service businesses have a visibility problem because their presence depends on the owner’s free time.

When the owner has capacity, the business posts, publishes, updates, and follows through. When delivery gets heavy, visibility disappears.

That creates Presence Debt.

The firm may still be doing excellent work, but the public evidence stops accumulating. The market cannot see the momentum. AI search cannot infer what is not present.

One agency publishes six strong articles in January, then goes quiet for eight months.

Another publishes one useful point of view every two weeks, updates case studies quarterly, appears in partner webinars, and keeps LinkedIn active with practical observations from client work.

The first had the bigger burst. The second has the stronger pattern.

That is Compound Presence: consistent, systematic visibility that accumulates over time.

Decision point: If your visibility plan only works when the owner has time this week, it is not a system. It is a recurring risk.

5. AI ignores what is not evidenced

AI search can ignore excellent work if the evidence is not public, structured, repeated, or connected to the right context.

It may ignore:

- referrals that happen only in private conversations - strong client outcomes with no case studies - niche expertise buried in sales calls - owner insights never turned into content - generic service pages that sound like every competitor - old content that no longer reflects the firm’s current position

A well-read colleague cannot recommend you for something they have never seen you demonstrate. AI search has the same limitation.

The corrected mental model

AI search does not recommend the company with the prettiest website or the cleverest AI tactic.

It recommends from the evidence it can find: consistent, credible, contextual signals that have accumulated over time.

Think less about optimizing for AI and more about becoming legible to a well-read colleague.

If that colleague could confidently explain what you do, who you help, and why you are credible — and the same evidence exists online — you are building Compound Presence.