Generative engine optimization

Be the answer AI gives your buyers.

Your buyers ask ChatGPT, Perplexity, and Google's AI Overviews before they ever visit a site. We build, manage, and optimize the GEO strategy that decides whether your brand is in those answers — and we measure it every week.

ChatGPT · Perplexity · Gemini · Claude · Google AI Overviews · Microsoft Copilot

AI visibility audit

Sample — illustrative

14%

you — share of answers

41%

top competitor

ChatGPT18% vs 44%
Perplexity22% vs 39%
Google AI Overviews9% vs 41%
Gemini11% vs 37%
Claude12% vs 43%

The kind of baseline every engagement starts with: how often each engine names you versus your top competitor, across the same set of buyer prompts.

Why now

Why a GEO plan matters more than ever.

Search is being answered, not listed. The shift is already under way — and the brands that plan for it now set the answers their competitors will have to fight.

The answer sits above the links

Google's AI Overviews and AI Mode summarize the answer before the first organic result. If your brand isn't in that summary, a page-one ranking gets far less attention than it used to.

Buyers ask assistants, not search bars

People now ask ChatGPT, Perplexity, Gemini, and Claude for recommendations and comparisons. They get one answer, not a page of options — and many never click through to a site at all.

An answer names a few brands, not ten

A results page shows ten links. An AI answer names a handful of brands and cites a few sources. There is no page two — you're in the answer or you're not.

Early movers compound

Engines keep citing the sources they already trust, and today's web becomes tomorrow's training data. The brands that shape their category's answers now get harder to displace every quarter.

The strategy

We build it, manage it, and keep optimizing it.

GEO isn't a one-time audit. It's a program — and we run all of it.

Build

We map the questions your buyers actually ask AI — category, comparison, “best X for Y,” and brand questions — then run them across every major engine to baseline where you're named, where competitors are, and which sources the engines trust.

DeliverableVisibility baseline, source map & GEO roadmap

Manage

We run the program: restructure key pages into clear, quotable answers, add schema markup and llms.txt, open access to AI crawlers, publish content for the prompts you're losing, and get you accurately represented on the third-party sites engines cite.

DeliverableFixes shipped in your CMS or codebase, every sprint

Optimize

Engines change constantly, so we re-run the full prompt set every week, track movement engine by engine, catch inaccurate answers early, and feed what's working back into the next round of content and outreach.

DeliverableWeekly tracking, monthly report & a standing fix queue

Measurement

What we measure, every week

Rankings don't describe an AI answer. These numbers do.

Mention rate

The share of buyer prompts where the answer names your brand.

Citation share

The share of cited links that point to your site, versus everyone else's.

Answer position

Whether you're the first recommendation or an afterthought when you are named.

Accuracy

Whether prices, features, and claims in the answer are right — wrong ones get fixed at the source.

Competitor share

The same numbers for every competitor you name, prompt by prompt.

Crawler access

Whether GPTBot, OAI-SearchBot, PerplexityBot, and ClaudeBot can actually reach your pages.

Engines we track

ChatGPT
Perplexity
Google AI Overviews
Google AI Mode
Gemini
Claude
Microsoft Copilot

Rates, not screenshots: AI answers change from one run to the next, so every prompt is sampled repeatedly and reported as a rate. You see real movement, not a lucky answer.

How it works

From baseline to a running program

Week 1

Baseline

We agree on the prompt set and competitors, audit crawler access and site structure, and run the first visibility baseline.

Weeks 2–3

Plan

Every prompt you're losing gets a cause — missing content, weak structure, or absence from the sources engines cite — ranked by impact.

Weeks 3–8

Execute

On-site fixes ship first. Content and third-party source work run alongside, on a sprint cadence.

Ongoing

Optimize

Weekly re-runs of the full prompt set, a monthly report, and a standing queue of what to fix next.

Proof

We run the playbook on our own site.

adres.ai is built the way we'll build yours. See for yourself at adres.ai/llms.txt.

190+

articles in the AdRes Journal, written to be quoted

200+

URLs in a sitemap that updates itself with every post

machine-readable guides for AI engines: llms.txt and llms-full.txt

schema.org types marked up across the site, from Organization to FAQPage

Operators, not an SEO shop. AdRes comes out of twelve-plus years inside the media and revenue operations of Apple, Hulu, Dentsu Aegis, Publicis Groupe, and Verizon Media. We build AI systems for a living — the same stack that runs our ad analytics — so we test how engines behave instead of guessing at it.

FAQ

Questions we hear about GEO

What is GEO?

Generative engine optimization is the work of getting your brand named, accurately, in the answers AI engines give — ChatGPT, Perplexity, Gemini, Claude, Microsoft Copilot, and Google's AI Overviews. SEO earns a ranked link; GEO earns a mention or a citation inside the answer itself.

How is GEO different from SEO?

GEO builds on SEO, because many AI engines search the web before they answer. But the goal and the measurement differ: an answer names only a few brands and cites a few sources, so we track how often you're named and cited across a set of real buyer prompts, not keyword rankings.

Why does GEO matter more now than a year ago?

More buying research now starts in an AI assistant or ends at an AI summary, before anyone sees a list of links. The sources engines trust today tend to keep getting cited, so the brands that build a GEO plan early are the hardest to displace later.

Can you guarantee we'll show up in ChatGPT?

No — and be wary of anyone who does. AI answers vary from run to run, and engines change often. What we commit to is measuring honestly, fixing the causes we find, and reporting the trend every week.

How long until we see movement?

Engines that search the web live, like Perplexity and ChatGPT search, can reflect changes within weeks. Answers drawn from a model's training data move more slowly, as new model versions are released.

Do we need to change our website?

Usually, yes. The most common fixes are structural: clear answers to the questions buyers ask, schema markup, and making sure AI crawlers aren't blocked. We make the changes in your CMS or codebase, or hand your team exact specs.

Find out what AI tells your buyers about you.

30 minutes, no pitch deck. We'll talk through your category, your competitors, and what a GEO program would look like for your brand.

Build · Manage · Optimize — measured weekly