GEO for Service Businesses: The 2026 Playbook

AI answer engines now sit between your customers and your website. When someone asks ChatGPT or Perplexity to recommend a law firm or a local service, the model returns one answer and cites a few sources. Getting your business into that answer is a discipline called Generative Engine Optimization (GEO), and the research shows it works: a Princeton-led study found that targeted GEO methods lift a page’s visibility in AI answers by up to 40 percent.

I build AI Search into every engagement. I grew a legal directory from zero to more than 100,000 monthly visitors and roughly 500 qualified leads a month, with AI Search workflows built in from launch, and it now earns citations in ChatGPT and Perplexity. This guide is the practical playbook I use.

What is generative engine optimization (GEO)?

Generative Engine Optimization (GEO) is the practice of structuring your content and your brand entity so that AI answer engines like ChatGPT, Perplexity, Google AI Overviews, and Gemini cite or recommend you inside the answers they generate. The goal is not a ranked link. The goal is to be part of the answer itself.

How is GEO different from SEO?

GEO extends SEO rather than replacing it, because pages that already rank are far more likely to be pulled into AI answers. A crawlable, fast, trustworthy site is still the foundation. What changes is the unit of visibility and the signals that decide who gets cited.

Dimension	Traditional SEO	GEO
Unit of visibility	A ranked URL	A cited passage inside the answer
What you optimize	Keywords, links, crawlability	Extractable answers, entities, statistics
Where you appear	The results page	Inside the generated answer
Primary signal	Backlinks and domain authority	Brand mentions and citation-worthiness
How you measure it	Rankings and clicks	Citations and share of AI answers

Why does GEO matter for service businesses right now?

Two things make this urgent for law firms and local businesses. First, most of your competitors have not started, so the citation window is wide open. Second, service businesses live on high-intent questions, and a single AI recommendation carries far more weight than a link buried on page one.

On the legal directory I built, AI Search workflows were part of the architecture from day one, including entity structure and structured data. It reached more than 100,000 monthly visitors and roughly 500 qualified leads a month, and it now appears in ChatGPT and Perplexity answers. You can read the full case study here.

The study behind most GEO advice is GEO: Generative Engine Optimization by Aggarwal et al., presented at ACM SIGKDD 2024. Its strongest method, adding quotations, improved visibility over the baseline by 41 percent on one measure and 28 percent on another. These are lab benchmark results, so treat them as direction rather than a promise.

Which GEO tactics actually move citations?

The tactics that move citations are attributable quotes, specific statistics, inline source citations, a front-loaded answer, a clean entity, earned mentions, and open AI search crawlers, with an llms.txt file as an optional extra. Each one costs little and compounds when stacked. The first three come from the Princeton research, which tested them in a lab benchmark. The rest come from what I see win in live projects.

• Add attributable quotes. In the Princeton GEO study, adding quotations was the strongest single method. It lifted a page’s visibility in AI answers by 41 percent on one measure and 28 percent on another. Quote a client, a named expert, or a primary source rather than paraphrasing.
• Add specific statistics. Statistics came a close second in the same study, at about 31 and 23 percent on the same two measures, because models treat numbers as extractable units. “437 reviews, a 4.9 rating, and an 89 dollar diagnostic fee” beats “great reviews and fair prices” every time. For law firms, the study lists Law and Government first among the categories where statistics helped most.
• Cite your sources inline. The study measured a 115 percent relative visibility gain for pages ranked fifth that added source citations, an equalizer effect that helps pages outside the top of the results. The same method cut visibility for pages ranked first by about 30 percent, so the gains go mostly to pages that are not yet on top.
• Front-load the answer. Answer the core question in the first 40 to 60 words, and make the first sentence under each heading read like a clean definition. AI retrieval weights opening content heavily.
• Make your entity unmistakable. Keep your name, address, and profiles consistent, and describe yourself with clear structured data (Person, Organization, and where relevant LocalBusiness). A well-linked entity is how a model confirms who you are.
• Publish an llms.txt file, as an optional extra. This is a short markdown map at your site root that gives AI systems a curated summary of your most important pages. It was proposed by Jeremy Howard at Answer.AI, and the spec lives at llmstxt.org. It costs little, but it is unproven: Google says you don’t need special AI text files to appear in its AI features.
• Earn third-party mentions. Independent analyses of AI citations in 2025 found that the large majority came from earned media rather than a brand’s own pages. Coverage, directories, and profiles that mention you matter more than raw backlinks here.
• Allow the AI search crawlers. Check each crawler for the job it does. For AI search, allow OAI-SearchBot (ChatGPT), Claude-SearchBot (Claude), PerplexityBot (Perplexity), Googlebot (Google and AI Overviews) and Bingbot (Copilot). GPTBot and ClaudeBot are training crawlers, so blocking them is a licensing decision and does not remove you from ChatGPT or Claude search.

How do I check whether AI engines already cite me?

Run your real customer questions through the AI engines and log who gets named. Paid trackers such as Ahrefs Brand Radar, Peec AI and Otterly now automate this. The manual method below costs nothing and takes about 20 minutes a month.

• Write 10 to 15 prompts a real customer would ask, such as “best [practice area] lawyer in [your city]” or “how do I choose a [your service]”.
• Run each prompt in ChatGPT, Perplexity, and Google AI Overviews.
• Log which businesses get named and cited, and whether you appear at all.
• Repeat monthly and watch your share of those answers move.

What does a 30-day GEO plan look like?

A 30-day GEO plan runs in four weekly steps: entity cleanup, money-page rewrites, an llms.txt file, and a citation baseline. Here is the version I use for a service business starting from a low base.

1. Week 1, entity cleanup. Fix your schema, keep your name and address consistent everywhere, and add your profiles as sameAs links so a model can confirm your identity.
2. Week 2, rewrite your money pages. Take your top three service pages and front-load the answer, add real statistics, and include one named quote each.
3. Week 3, check crawlers and add llms.txt. Confirm the AI search crawlers above are allowed in robots.txt, then add an llms.txt file at your root.
4. Week 4, baseline and seed. Run your first citation audit across the three engines, then seed two or three earned mentions on directories and relevant sites.

If you would rather have this done for you, that is exactly what my AI Search optimization service covers, from entity architecture to the citation audits above.

Frequently asked questions

What is generative engine optimization (GEO)?

GEO is the practice of structuring your content and brand entity so that AI answer engines like ChatGPT, Perplexity, and Google AI Overviews cite or recommend you inside the answers they generate, rather than only ranking your page as a link.

Is GEO different from SEO?

GEO extends SEO rather than replacing it. Strong technical SEO remains the foundation, because pages that already rank are more likely to be cited. GEO adds a layer focused on how AI engines select and reuse content.

How long does GEO take to show results?

On-page tactics such as adding statistics and cleaning up structure can show measurable citation lift within two to four weeks of a crawl. Entity and earned-media signals compound over months. A realistic expectation from a low base is meaningful movement within a quarter.

What is the single most effective GEO tactic?

Adding quotations from named, credible sources. In the Princeton GEO study it was the strongest single method, lifting visibility by 41 percent on one measure and 28 percent on another. Adding specific statistics came a close second, at about 31 and 23 percent. These are lab benchmark results, so test them on your own pages.

Should I block AI crawlers like GPTBot?

Blocking GPTBot does not remove you from ChatGPT search. GPTBot and ClaudeBot are training crawlers, so blocking them is a licensing decision. For AI search, allow OAI-SearchBot (ChatGPT), Claude-SearchBot (Claude), PerplexityBot (Perplexity), Googlebot (Google and AI Overviews) and Bingbot (Copilot). Google-Extended controls use by Gemini models, not your place in Google Search.

How do I start earning AI citations?

Start by moving before your competitors do, since most of them have not. If you want a review of where you stand today, book an intro call and I will run a citation audit on your site beforehand and bring specifics. You can also find more guides from the field in the resources library.