Who gets recommended when a founder asks an assistant for a consultant

The questions an operator types before contacting a consultancy

The question names a problem, an industry and a size before it names a firm. "pricing consultant for a b2b software company with 40 employees", "operations consultant in Chicago who has worked with food manufacturers", "consultant to help a canadian company enter the us market", "small consulting firm for a family business succession, not one of the big firms". Then, about one firm: "who would actually do the work", "do they take companies our size", "how do they charge, project or retainer".

The person asking is often the one who will pay, and they are not asking for a framework. They are asking who has solved this problem for a company like theirs, and whether the people on the website are the people who will show up. An AI assistant answers by fetching pages and reading for those facts, and it gives an AI recommendation from the pages that state them.

Consulting sites are written for a different reader: a procurement committee that already knows the firm and wants reassurance. The words are "transformation", "value creation" and "end-to-end", which match no question a founder types at eleven at night. The site that gets named is the one that reads like the answer to the founder's question.

What a consulting firm's site controls

Four things. Whether the niche is stated in the words a buyer uses: the problem, the industry, the company size. Whether the people are on the page as people, with what they did before and what they do now. Whether the evidence is real and readable: engagements described in text, publications the firm wrote, clients who agreed to be named. And whether the firm reads as one entity across its pages and the partners' profiles elsewhere.

Niche is the match. "We help business-to-business software companies with 20 to 200 employees set and change their prices" is a sentence that matches the first question above almost word for word. "Growth strategy for technology companies" matches it loosely. "Strategy consulting" matches everything and therefore nothing. A firm does not have to narrow what it does to be matched; it has to state each thing it does in its own sentence, with the industry and the size in it.

People are the second match, and one consulting sites often leave out. A buyer asking "who would do the work" is matched to a page that names the people, says what they did before consulting, and says which engagements they lead. A page of stock photographs and a paragraph about "our team of experienced professionals" gives the assistant no one to name, and a firm that names its people gives it someone.

What the firm's site has to state

Each row is a buyer's question, the fact about the firm that answers it, and where it belongs. Everything in the middle column is a statement about the firm, and the engagements described are real ones or none.

A consultancy is a business at an address even when the work is remote, so the markup is LocalBusiness schema, using schema.org's LocalBusiness type itself because its ProfessionalService subtype is marked deprecated, or Organization for a firm with no office at all, carrying the name, address and telephone that Google's local business documentation lists. Each offer is a Service node with serviceType in the buyer's words and audience naming the industry and size. Each consultant is a Person node. All of it as JSON-LD, repeating what the visible page says.

Named engagements, anonymized engagements, and what never to invent

A named engagement is the strongest evidence a consulting site can carry, and it exists only with the client's written agreement to be named and to the description. Without that, the engagement is described without the name: the industry, the company's size, the problem as it was posed, what the firm did, how long it took. Those are facts a buyer can match to their own situation, and an assistant can quote them.

Results are where the line is. A result stated on your site is a claim about someone else's business, and it stands only if the client confirmed it in writing. Absent that, state what was done and let the client's own words, quoted with permission, say how it went. Do not invent an engagement, a client or a figure to fill a gap. An assistant that names you for a claim you cannot support has done you no favor, and a buyer who checks will find out.

Publications are evidence of a different kind, and consultancies underuse them. A piece of your own research with its method stated, a data set you collected, a plain explanation of how a problem is usually solved: these are pages an assistant can cite as sources, not only as a name in a list, which is what citability means, and a body of them on one subject is what topical authority means. The one controlled study we know of, Aggarwal and colleagues' "GEO: Generative Engine Optimization" (KDD 2024), found that adding citations, quotations and statistics to a source page raised its visibility in generated answers by up to about 40% on their benchmark. Their benchmark was not consulting sites; the mechanism is what carries over.

What our crawl found

In our audit of 9,483 small business websites across 40 trades and 90 metros in the US, Canada, Europe and the UAE, run between late August and early September 2026, 61.5% had no question-and-answer content of any kind, and 50.7% had at least one heading phrased as a question with nothing under it a machine could isolate. A rhetorical question heading ("Ready to transform?") counts toward that 50.7%; an answered one ("How long does a pricing project take?") does not.

17.8% carried no schema.org markup at all, and most of the rest described the website rather than the business. A consultancy with an Organization node, a Service node per offer and a Person node per consultant has stated in machine-readable form what its prose usually leaves implied. And 43.1% were inconsistent about who and where they are across their own three main pages; for a firm of three partners working from home, the address that appears can be a former office, a registered agent, or nothing.

What to fix first, in order

1. Settle the identity: one firm name, one address or a stated remote base, one main contact, identical on every page and in every profile the partners keep.
2. Write one page per offer in the buyer's words: the problem, the industry, the size, the engagement model, the typical duration, the fee model. Each buyer question as a heading, the answer under it. This is the answer page for that offer.
3. Put the people on the page as people: name, what they did before, what they lead now, where they are based, a photograph of them and not of a skyline.
4. Describe the work: named engagements with written permission, anonymized ones by industry, size, problem and what was done. No results without the client's confirmation. No invented cases.
5. Add the markup as JSON-LD: a LocalBusiness or Organization node, a Service node per offer with audience, a Person node per consultant, all matching the visible text. Google's introduction to structured data says the markup describes what a visitor can see.
6. Publish something you measured, with the method. Then open the offer pages with scripts off, confirm the text is in the HTML, and run the crawler access check.

What not to do

Do not write testimonials yourself or present your own words as a client's; Google's review snippet documentation excludes reviews that are not based on a genuine experience. Do not put a client's logo on the page without permission.

Do not stuff pages with "management consulting Chicago, Chicago strategy consultants, consulting firm Chicago". Google does not read the keywords meta tag, as its list of supported meta tags shows, and an assistant reading prose wants the sentence with the industry and the size in it.

Do not promise outcomes, and do not accept a promise of placement. What an assistant says is decided inside the company that runs it, and no agency, ours included, has a way in. What can be measured is how often your firm is named on a fixed panel of buyer questions for your offers; how we measure it is written down in full.