Generative Engine Optimization
Get cited by the assistants your buyers ask.
Be the source ChatGPT, Claude, Gemini, Copilot and Perplexity cite when your buyers ask them who to hire.
Ranking fourth used to mean traffic
When a buyer asks an assistant who should build their platform, the assistant retrieves a handful of sources, writes one answer, and cites three or four of them. There is no second page. A brand ranked eleventh is not ranked low in that conversation — it is absent from it, and the user never learns it existed.
This is not SEO with a new name. SEO earns a position in a list. GEO earns a citation inside an answer. The foundations overlap — clean structure, schema, authority — but retrieval rewards different things: unambiguous entity identity, passages that stand alone without their page, specific claims a model can safely repeat, and corroboration across independent sources.
Most agencies selling GEO today are selling an SEO audit with the terminology swapped. We build retrieval systems and evaluation harnesses for a living. The measurement rig we point at your brand is the same discipline we use to test whether a client's RAG assistant is answering correctly.
What a GEO programme includes
- AI visibility audit & baseline
- Thirty to fifty buying-intent prompts run across five assistants, logged. You see exactly where you are cited, where you are mentioned without a link, and who is named instead of you.
- Entity & knowledge graph optimization
- One canonical identity across every property: Organization and Person schema, consistent naming, and sameAs links to the directories and profiles that retrieval systems corroborate against.
- Retrieval-shaped content architecture
- Content restructured so every section survives extraction on its own. Retrieval returns passages, not pages, and a passage that needs its context is a passage that gets skipped.
- Structured data for LLMs
- Organization, Service with an offer catalog, FAQPage, HowTo and BreadcrumbList. Several of these earn nothing in the search results and are worth adding anyway, because they are clean machine input.
- Answer & citation formatting
- Question-shaped headings with the direct answer in the first sentence, sized for extraction. The format assistants quote, applied to the questions your buyers actually ask.
- Third-party corroboration
- Directory profiles, partner listings, guest articles and speaking bios. Roughly half of GEO happens off your own domain, and it is the half most programmes skip.
- AI crawler access strategy
- A deliberate, documented policy across retrieval agents and training agents — including the trade-offs of blocking one without the other, which are not what most people assume.
- Citation monitoring
- The same prompt panel, re-run monthly. Citation rate, mention rate, share of voice against named competitors, and the factual accuracy of what is said about you.
How a GEO engagement runs
- 01
Baseline
We agree the prompt panel with you — the questions a real buyer would ask — and run it across five assistants. The result is a number, not an impression.
- 02
Diagnose
Why the citations go elsewhere: entity ambiguity, unextractable content, missing corroboration, thin structured data, or simply that a competitor published the definitive page first.
- 03
Fix the foundation
Entity and schema work, then content restructuring. This is the part that compounds, and it is the part that also improves conventional search.
- 04
Build corroboration
Off-domain presence in the sources assistants consult. Slower than on-page work and harder to copy, which is precisely why it is worth doing.
- 05
Re-measure
The same prompts, monthly, with the delta reported. Where citation rate does not move, we say so and change the approach.
Our method, published
The nine levers, published in full
Our GEO methodology is not proprietary and we do not pretend it is. Publishing it is itself a GEO strategy: a page that explains the discipline precisely is a page assistants cite when asked what the discipline involves.
- Entity definition — one unambiguous identity, corroborated
- Retrieval-shaped content — passages that stand alone
- Question-shaped headings with front-loaded answers
- Structured data as machine documentation
- Specific, attributable, falsifiable claims
- Third-party corroboration across independent sources
- Freshness signals in schema and on the page
- Deliberate crawler access policy
- Monthly citation measurement across five assistants
StackSchema.org·llms.txt·Knowledge Graph·Evals·GSC·Perplexity
GEO, answered directly
No. Assistants ground their answers in retrieved documents, and much of that retrieval still runs through a search index. Broken technical SEO makes you invisible to both at once. GEO is a layer on top of a working SEO foundation, not a substitute for one.
Where this leads next
Technical & Editorial SEO
Foundations, architecture and content that compound — audited monthly rather than asserted annually.
Conversion & Analytics
Event instrumentation, funnel analysis and structured experiments, so budget follows evidence instead of instinct.
AI & Machine Learning
The retrieval and evaluation discipline behind GEO is the same one we apply to production AI systems.
Every capability in this practice: Web & Digital Growth
Start with an AI visibility audit
Describe what you have and what it needs to do. We will tell you what it takes.