A CFO search is coming up at a healthcare services company. The CEO asks Claude which executive search firms specialize in finance leaders for healthcare. A few firms get named. Every other firm is absent from that conversation.
After we restructured RE/MAX Commercial Advisors Group's site content for AI answer engines in 2026, Patrick J. Buckhoff, CCIM, who leads the brokerage, reported inbound calls from people who found the firm through AI tools. More on that below. First, the answer you came for.
To get cited by ChatGPT, Claude, and Perplexity, three things matter more than everything else: specificity, structured data, and provenance. Say exactly who you serve, what you do, where, and what happened. Make the page easy for a machine to read. Show where every fact came from. Do those three on the right pages and you've done most of the work of generative engine optimization.
What makes a website citable by AI?
A website is citable by AI when it contains specific, checkable sentences a model can repeat with confidence, structured so the model can find them, and backed by sources, names, and dates that show the facts are real. Give a model a sentence worth lifting and a reason to believe it, and you get cited.
That maps to the CITE Framework we use with clients: Claim, Instrument, Testimony, Evidence. Specificity is the Claim. Structured data is the Instrument. Provenance runs through Testimony and Evidence.
Specificity: give the model a sentence it can repeat
When a buyer asks ChatGPT, "Who are the best sell-side M&A advisors for HVAC companies in Texas?" the model needs a sentence on someone's website that answers that question. If your site says something a thousand firms could say, the model has nothing to work with. It moves on to the firm that wrote the sentence.
Here's a homepage line we see constantly:
Before: "We deliver tailored strategic solutions that help businesses achieve their goals."
A model can't repeat that, because it's true of everyone.
After: "We run sell-side M&A for founder-owned HVAC and mechanical services companies between $10M and $50M in revenue in Texas and Oklahoma."
Now there's an industry, a deal type, a size band, an ownership type, and a geography. A model can match that to a buyer's question and say it back.
Same move on a service page:
Before: "Our experienced team provides comprehensive executive search services across a variety of industries."
After: "We recruit CFOs and VPs of Finance for private-equity-backed healthcare services companies, and 80% of our searches close within 120 days."
(That second one is a hypothetical. Only publish a number like that if it's true and you can stand behind it.)
The research backs this up. The Princeton and IIT Delhi GEO study (Aggarwal et al., KDD 2024) found GEO methods can boost visibility by up to 40% in generative engine responses. Adding quotations, statistics, and source citations were among the top-performing methods. Keyword stuffing performed worse than doing nothing on the study's main visibility measure.
The models reward the same specifics a careful buyer does.
Where specificity matters most
Start with the citable core of your site:
- Your about page. Who founded the firm, when, with what background, serving whom, where.
- Your core service pages. One page per service. Who it's for, what's included, what the outcome looks like.
- Your homepage. One sentence near the top that says who you serve and what you do. Plain words.
- Your case studies. Named client, the problem, what you did, the result as a number.
- Your people pages. Named humans with real credentials.
When a buyer asks an AI tool to recommend a firm, those are the pages the model needs.
Structured data: make the page easy for a machine to read
A specific sentence buried in a slider or locked inside an image is still invisible. Structure is how the machine finds it.
Structured data covers more than schema markup. It's the whole shape of the page.
Use question headings with direct answers
Write headings the way buyers ask questions. "Who do we work with?" "How does a sell-side process work?" "How long does an executive search take?" Then put a two to four sentence answer directly underneath. Lead with the answer. Add detail after.
That pattern hands an AI tool a clean question and answer pair it can lift whole. We go deeper on which formats get quoted in the content patterns LLMs quote.
Add schema that matches the visible text
Schema is code that labels what's on your page: this is an organization, this is a person, this is a service, these are FAQs. For a B2B service firm, the core set is Organization (or ProfessionalService), Person, Service, Article, and FAQPage.
One rule matters above the rest. Your schema must say the same thing your page says. Google Search Central's guidance on AI features states that no special schema, AI text files, or markup is required to appear in AI Overviews or AI Mode, and that standard SEO applies, including "making sure your structured data matches the visible text on the page."
Do the standard schema well and keep it honest. The full breakdown by type is in structured data for answer engines.
Let the AI crawlers in
Open yoursite.com/robots.txt. Look for these user agents:
- OpenAI: OAI-SearchBot and ChatGPT-User (search and user requests), GPTBot (training)
- Anthropic: Claude-SearchBot and Claude-User (search and user requests), ClaudeBot (training)
- Perplexity: PerplexityBot and Perplexity-User
The search and user agents are the ones that matter most for citations. If any of them sit under a Disallow: / line, that may limit whether those tools can retrieve and cite your pages. We regularly find these blocks on service-firm sites, often added by a security plugin or a developer being cautious, with nobody on the firm side aware of it. If you want to be cited, allow them.
Render your content in HTML
Some sites load their text with JavaScript after the page opens. Many crawlers don't wait, so they see an empty shell.
Quick test: right-click your service page and choose "View page source." If you can read your actual service description in that source code, you're fine. If you see a pile of scripts and no words, your most important copy may be invisible to the tools you want citing you. Ask your developer about server-side rendering or static HTML for core pages.
Provenance: show where every fact came from
Provenance is what makes a specific claim believable.
Buyers and models both trust what they can verify. A claim with a name, a date, a source, and a credential attached is one a model can repeat with confidence.
Name your people and their credentials
"Our experienced team" is a ghost. Swap it for real people.
Before: "Our team brings decades of combined experience."
After: "Maria Torres, CPA, spent 14 years as a tax partner at a regional firm before founding the practice in 2012. She leads all engagements for engineering and architecture clients."
(Hypothetical. Use your real people.)
Give every key person a bio page with credentials and niche, byline articles with those names, and add matching Person schema.
Publish case studies with numbers
A case study with a named client and a percentage is one of the strongest citable assets a service firm can own.
Before: "We helped a leading firm grow its online presence."
After: "Travel Again Advisory, a travel-industry M&A advisory, saw a 1,011% increase in organic search visits year over year after we rebuilt its site for Google and for AI answer engines."
That second version is real. It's from our Travel Again Advisory case study, and the firm also landed a sell-side client who found it on Google and hired it.
Cite your sources and date your content
When you use a number, link to where it came from. When you publish an article, put a visible date and a byline on it. When you update a page, change the date.
This is the Evidence letter in the CITE Framework, and buyers trust it for the same reasons machines do.
Get other sources talking about you
Your own site is one witness. Podcast appearances, industry association listings, directory profiles, press mentions, and client reviews all tell a model that the claims on your site hold up.
Keep your description consistent everywhere. Two different stories give a model no confident sentence to repeat.
The proof: RE/MAX Commercial Advisors Group
RE/MAX Commercial Advisors Group is a Houston commercial real estate brokerage led by Patrick J. Buckhoff, CCIM. In 2026 we restructured the firm's site content for AI answer engines.
After we restructured the site, Patrick reported: "Recently I've gotten three calls inbound that people found us online. Two from chat, one from Claude."
Inbound calls that name an AI tool are the scoreboard that counts. Patrick also notes, "We get a fair amount of inbound leads every week." That's coming from a firm that had its best year yet in 2024, coming out of the toughest commercial market since 2008. The full story is in the RE/MAX Commercial Advisors Group case study.
Why this is worth doing now
Buyers have already moved. Responsive's October 2025 survey of 350 B2B buyers found generative AI has overtaken traditional search for a quarter of B2B buyers, and nearly two-thirds use it as much as or more than search when researching vendors.
If the shortlist forms inside ChatGPT, your website has to be in the room when it forms.
A worked example: one afternoon of rewrites
Say you run a five-partner executive search firm in Denver that places operations leaders in manufacturing. This is a hypothetical.
Your homepage headline says "Connecting Exceptional Talent with Exceptional Organizations." Your about page talks about passion and partnership. Your service page lists "retained search, contingency search, and leadership advisory" with no industries named. Your robots.txt blocks OAI-SearchBot and Claude-SearchBot because a plugin did it two years ago.
The afternoon:
- Rewrite the homepage line to: "We recruit plant managers, VPs of Operations, and COOs for mid-sized manufacturers across the Mountain West."
- Rewrite the about page with each partner's name, background, and the year the firm started.
- Add three question headings to the service page ("What roles do we recruit for?" "Which industries do we serve?" "How long does a search take?") with short direct answers.
- Add Organization, Person, and FAQPage schema that matches the visible text.
- Remove the AI crawler blocks.
- Publish one case study with a named client and a percentage result.
All of it makes the firm easier to find, read, and believe.
When you're done, check your work with a quick prompt test, and keep going with the full generative engine optimization playbook.
Frequently asked questions
How do I get my website cited by ChatGPT? Make specific, checkable claims about who you serve, what you do, where, and what happened, then structure those pages so a machine can read them and back them with named sources and credentials. Allow AI search crawlers like OAI-SearchBot and Claude-SearchBot in your robots.txt and make sure your content renders in plain HTML. Specificity, structured data, and provenance do most of the work.
Which pages on a service firm's site matter most for AI citations? Your about page and core service pages. They describe who you are and what you do, which is exactly what a model needs when a buyer asks for a recommendation. Case studies with named clients and numbers come next, followed by bios for the people who do the work.
Should I block or allow AI crawlers in robots.txt? If you want to be cited, allow them. The agents that matter most for AI search are OAI-SearchBot and ChatGPT-User (OpenAI), Claude-SearchBot and Claude-User (Anthropic), and PerplexityBot and Perplexity-User (Perplexity); GPTBot and ClaudeBot are the training crawlers. A blanket block, often added by a security plugin or a cautious developer, can keep you out of AI search results.
Do I need special AI schema or an llms.txt file to get cited? Google says no special schema, AI text files, or markup is required to appear in AI Overviews or AI Mode, and standard SEO applies. Standard schema such as Organization, Person, Service, and FAQPage still helps every engine understand your pages, as long as it matches the visible text.
How long does it take to start getting cited by AI tools? It varies by how often the tools refresh and how much proof already exists about your firm elsewhere on the web. Rewriting your core pages is a few weeks of work. Building the third-party mentions that make a model confident in you takes longer and keeps compounding.
If you want to see how your site reads to ChatGPT, Claude, and Perplexity today, we'll walk through it with you.