Consulting AIO: How Firms Get Named in AI Answers

Consulting AIO (AI Optimization) is getting your firm named when a buyer asks ChatGPT, Gemini, Perplexity, or Claude who can solve a problem. It comes down to three things: a narrow specialty stated the same way everywhere, case studies and expertise pages that assistants can read and quote, and third-party sources that name your people as experts.

How do AI assistants decide which consultants to name?

AI assistants draw on two sources, and each one needs a different fix. When an assistant answers from what it learned in training, it repeats the firms and people that showed up most consistently in that text: trade press, conference programs, rankings, forum threads, books. When it searches the web first, it pulls a handful of pages that match the question and builds the answer from them. ChatGPT, Gemini, Claude, Perplexity, and Microsoft Copilot can all search this way. Google AI Overviews and AI Mode build answers from Google's index.

Influencing the training side is GEO (Generative Engine Optimization). Getting retrieved and cited is AEO (Answer Engine Optimization). Consulting firms struggle with both for a structural reason. Software companies have review sites and comparison articles that assistants retrieve constantly. Most consultancies have service pages that say they "partner with clients to drive lasting transformation," and very little independent text that says what they are specifically good at.

Picture a buyer asking who can help a regional hospital system reduce readmissions. An assistant has nothing to match against transformation language. So it names the large generalist firms it saw everywhere in training, or the boutique that published the one clear, readable page about readmission programs. You want to be that boutique.

Decide what you want to be named for

The first decision is the phrase you want an assistant to attach to your firm, and it should be narrower than feels comfortable. Buyers rarely ask for "a strategy consultancy." They ask for help with post-merger integration at regional banks, pricing for usage-based software, or safety compliance programs for construction contractors.

Compare two firms. One describes itself as an operations and growth advisory. The other says, on every partner bio, conference profile, and byline, that it fixes warehouse labor planning for third-party logistics providers. The second firm gives a model a consistent pattern to learn and gives a retrieval system pages that match real questions. The first competes with every firm on earth for a phrase nobody types.

This matters for mechanical reasons. During training, a model connects a name to a topic when many independent sources repeat that pairing. During retrieval, the pages that answer the exact question get pulled. A broad description fails on both counts.

The common mistake is listing a dozen practice areas with equal weight. Pick one to three you want to own, write about those relentlessly, and let the other services exist on the site without competing for the same language.

Which questions should your firm track?

Track the questions buyers actually type, collected from discovery calls, RFPs, and the first email a prospect sends. They fall into two groups. Vendor questions ("Who are good consultants for ERP selection at mid-market manufacturers?") are where you want your firm named. Problem questions ("How should a mid-market manufacturer run an ERP selection?") are where you want your content cited as the source.

Problem questions deserve more attention than most firms give them. A buyer who reads an answer built on your framework, with your page cited underneath, often reaches out before ever asking for a vendor list.

Run each question several times on each assistant and vary the wording. AI answers are non-deterministic, so the same prompt can return a different shortlist an hour later. A single check tells you almost nothing. What you want is how often your firm appears across many sampled answers and which pages get cited, whether you appear or not. nonBot AI runs these questions across ChatGPT, Gemini, Perplexity, Claude, and Google AI Overviews and shows which pages each answer cites. For a short list, a spreadsheet works: ask, note the firms named and the URLs cited, repeat.

The cited URLs are the most useful thing you will collect. If Perplexity keeps citing a trade association guide and a competitor's blog post for your core question, you now know the two pages you need to outdo or get mentioned in.

How do you write case studies AI can retrieve and cite?

Write each case study as its own web page that names the client type, the problem, the method, and the outcome in plain sentences an assistant could quote. Most consulting case studies miss all four. "A leading global company faced complex challenges, and we delivered a tailored solution that created significant value" matches no question a buyer would ask.

A problem-led page looks like this:

• Title as the problem: "Cutting claims processing time for a regional health insurer" instead of "Client Success Story."

• Opening summary: two or three sentences covering the situation, what you did, and what changed, written so the paragraph makes sense quoted alone.

• Specific outcomes: real figures where the client permits, described in the units the client tracked.

• Named lead: the partner who ran the engagement, linked to a bio that states the same specialty.

• Attributed quote: a client quote with a role and organization type, even if the name stays private.

Confidentiality is the usual objection, and it is solvable. "A regional health insurer covering two states, with a backlog after a systems migration" protects the client and still gives a retrieval system the nouns it needs to match a question. Specific description does most of the work. The client's brand name mostly adds credibility on top.

Ungate the thinking you want quoted

Consulting firms habitually put their best thinking in gated white papers and PDF reports. Crawlers cannot fill out forms, so that thinking is invisible to every assistant that searches, and it never enters training data from your site. Publish the core argument as an open web page and keep a downloadable version with extra material for lead capture.

Then check whether AI crawlers can reach your site at all. Open your robots.txt file and look for rules covering OpenAI's GPTBot and OAI-SearchBot, Anthropic's ClaudeBot, PerplexityBot, and Google-Extended, which governs whether Google can use your content for Gemini training. Blocking search crawlers removes you from retrieval. Blocking training crawlers is a legitimate choice, but it means future models learn about you only from other people's sites.

Also check that your text appears in the page source. Sites built so that content loads through JavaScript after the page opens can look empty to crawlers that do not render scripts.

How do you get your expertise into AI training data?

Get named by independent sources that cover your field, because that is the text models learn associations from. Your own site counts, but a single source repeating itself is weak evidence. The sources that build a durable association for consultants include:

• Trade publications in your vertical, through bylined columns and journalist quotes

• Industry association reports and conference session pages

• Podcasts that publish full transcripts

• Books and research co-authored with clients or universities

• Practitioner communities on Reddit and Quora where your buyers ask for advice

Picture a supply chain partner who writes a regular column for a logistics trade magazine, speaks at an industry conference whose session page describes her talk, and gets quoted in coverage of warehouse labor shortages. Each of those pages ties her name, her firm, and her topic together, and each one lives on a site the firm does not control. Repeated over time, that pattern is what a model can pick up.

Be realistic about the timeline. Nobody outside the AI labs can see exactly which documents went into a given model, and retraining happens on schedules you do not control. Treat this as long-term work that also pays off in retrieval along the way, since trade press pages get cited by assistants that search.

Wikipedia deserves a note. If your firm has enough independent press coverage to meet Wikipedia's notability standards, an accurate article helps. Writing your own article usually ends with deletion and a conflict-of-interest flag.

What trust signals do AI answers pick up for professional services?

The signals that carry weight are the ones a careful human researcher would check: who your people are, what they have published, who vouches for them, and whether those details agree across sources. Assistants pick these up indirectly, through the sources they learned from and the pages they retrieve.

• Named experts with consistent bios: each partner's specialty phrased the same way on your site, LinkedIn, speaker pages, and author boxes.

• Credentials stated plainly: CPA, bar admissions, board certifications, and named prior roles, written as text rather than logos.

• Consistent firm facts: the same firm name, locations, and practice areas across your site, directories, and Google Business Profile.

• Third-party rankings and reviews: the directories your buyers already trust in your category.

• Structured data: Person and Organization schema help search engines understand who is who, which matters because AI search features retrieve from those indexes. Schema alone will not get a page cited.

The third-party sources differ by type of firm:

Mistakes that cost consulting firms visibility

Most lost visibility comes from a few habits that feel productive.

• Reacting to a single prompt: checking one question each morning and rewriting pages when the answer changes. The variation is normal. Judge patterns across many runs.

• Buying listicle placements: paying to appear in thin "top consultants in Chicago" pages. They may get retrieved for a while, but they carry little weight and read like the paid placements they are.

• Renaming practice areas: every rebrand of a service line breaks the association you spent years building in third-party text.

• Conflicting bios: a partner described as a digital transformation leader on one site and a pricing expert on another gives models two weak signals instead of one strong one.

• Treating AIO as a website project: the site matters for retrieval, but most of the training-side signal lives on pages you do not own.

What to do this week

Start with the questions, because every later decision depends on them.

1. Pull notes from your last ten discovery calls and copy each buyer's opening question word for word.

2. Run those questions in ChatGPT, Perplexity, and Google a few times each, and record which firms get named and which URLs get cited.

3. Pick the case study closest to the most common question and rewrite it as an open, problem-led page with a quotable opening summary.

4. Check robots.txt for the AI crawlers listed above and confirm your best thinking is readable without a form.

Before Friday, do the first step: open your call notes and paste ten buyer questions into a spreadsheet.

FAQ

How long does consulting AIO work take to show up in AI answers?

Retrieval changes can appear fairly quickly. Once a new page is indexed and matches a question well, assistants that search can start citing it. Training changes take much longer because they depend on when models are retrained and what goes into the data, neither of which you control.

Does AI visibility matter if most of our work comes from referrals?

Yes, because referred buyers still check. A peer recommends your firm, and the buyer asks an assistant about you or about the problem before calling. The answer either supports the referral or hands them three alternatives.

Can we tell when an AI assistant sent us a client?

Partly. When someone clicks a cited link, the visit shows up as referral traffic from domains like chatgpt.com and perplexity.ai in your analytics. Many buyers read the answer and then search your name or email a partner directly, so add an AI assistant option to your intake form's "how did you hear about us" question.

Do partner LinkedIn posts help with AI visibility?

They help buyers find and judge your partners, but their effect on AI answers is uncertain. Assistants cite open web pages far more readily than LinkedIn posts, and how much LinkedIn content reaches training data is not public. Turn your strongest posts into articles on your site and pitches to trade publications.