Generative Engine Optimisation (GEO)

Generative Engine Optimisation (GEO)

Earn the citations that decide what AI engines say about you

Generative engine optimisation, or GEO, is the practice of improving how AI engines such as ChatGPT, Claude, Gemini, Copilot, Perplexity, Google AI Mode, and Google AI Overviews represent a company when they answer a buyer's question. Answer engine optimisation, or AEO, is the same discipline named for its destination: the answer itself. Third Hemisphere delivers GEO as a PR-led service, because the engines compose their answers from published sources, and earned media is the source they weight most heavily. This page explains the discipline. The Fourth Hemisphere: Marketing for AI is the full program it sits inside.

What GEO and AEO actually mean

An AI engine answering a buyer's question composes one response from the sources it trusts, and it cites a handful of them. GEO is the work of becoming one of those sources: publishing content an engine can retrieve and quote, keeping your entity data consistent everywhere the engine reads about you, and earning coverage in the outlets the engine treats as authoritative. AEO describes the same work from the buyer's side, because the deliverable is the answer they are given. A provider quoting them as separate services with separate prices is selling one job twice.

Why GEO is a PR problem before it is a technical one

The technical half of GEO, structured data, retrievable content, and consistent entity records, is a precondition. The half that moves the answer is what independent sources say about you, because the engines weight third-party authority over anything you publish about yourself. That is earned media, and earning it needs journalists to say yes. Third Hemisphere has placed over 20,000 stories for over 150 clients since 2015, which is why we deliver GEO as a communications discipline with the technical work built in, and why an SEO vendor without a newsroom can only deliver the precondition.

PACT: how we measure the answer

Every GEO engagement runs on the PACT diagnostic. We ask the seven engines, ChatGPT, Claude, Gemini, Copilot, Perplexity, Google AI Mode, and Google AI Overviews, 50 real questions your buyers ask, record every answer and citation, and score the results across Presence, Advocacy, Consensus, and Truth. The score tells you whether you appear, whether the engines recommend you or merely list you, whether they agree about you, and whether what they say is accurate. It also names the competitor holding your slot and the sources the engines drew on, which is the work plan: change those sources and the answer follows.

What the work contains

A GEO program opens with the PACT diagnostic, fixes what is cheap to fix, then builds the citable record. Corrections come first: inaccurate entity data, missing structure, and third parties carrying wrong information move answers faster than new authority does. The earning phase follows, placing your story, your data, and your executives' commentary in the outlets the engines cite. Quarterly re-audits re-run the same 50 questions, so progress is measured in the answers themselves rather than in proxy metrics.

Where GEO ends and Marketing for AI begins

GEO is the discipline: measure the answers, correct the record, earn the citations. The Fourth Hemisphere: Marketing for AI is the strategy layer above it. That program takes your whole marketing strategy as its input, rates every area of it for AI visibility, and directs the specialists who deliver each part, including any SEO or digital agency you already retain. Start here if you want the discipline delivered. Start with Marketing for AI if you want your entire plan read against what the engines do.

  • The PACT diagnostic: 50 buyer questions across the seven engines
  • Scoring across Presence, Advocacy, Consensus, and Truth
  • Entity consistency across your site, profiles, and third-party databases
  • Structured, retrievable content built for engine citation
  • Earned media in the outlets the engines treat as authoritative
  • Factual corrections pursued at the source
  • Quarterly re-audits that measure movement in the answers themselves
  • A defined hand-off to The Fourth Hemisphere: Marketing for AI when the whole strategy needs directing

Our GEO Method

  1. PACT Diagnostic We ask the seven engines 50 real questions your buyers ask, record every answer, citation, and omission, and score the results across Presence, Advocacy, Consensus, and Truth. You see where you appear, who is named instead, and which sources the engines relied on.
  2. Correct We fix what is cheap to fix first: inconsistent entity data, missing structured data, content a model cannot quote without the whole page, and third parties carrying wrong information about you.
  3. Earn We build the citable record: earned media in the outlets the engines weight, bylined commentary from your executives, and research designed to be cited. This is the half that needs journalists to say yes, and it is the half this agency was built on.
  4. Re-audit Each quarter we re-run the same 50 questions and re-take the PACT score, so you see the answers changing question by question rather than trusting a report that describes activity.

Who this is for

  • Companies whose category questions are answered by AI engines without mentioning them
  • Marketing leaders asked by their board what GEO or AEO is and whether it is funded
  • Companies whose AI answers misstate their products, leadership, or record
  • Teams already running SEO who need the earned-media half an SEO agency cannot deliver
  • Brands with strong media coverage that has not translated into AI-engine presence

What this delivers

One Discipline, Honestly Priced

GEO, AEO, AI SEO, and LLM SEO are one job under four names. We define the work, price it once, and publish the method, so you can compare providers on deliverables rather than vocabulary.

Measured in the Answers

The PACT diagnostic gives you a baseline score and a question-level record before any work begins, and quarterly re-audits re-take the same measure. Progress is visible in what the engines actually say.

The Earned-Media Half, Delivered

The engines weight independent, authoritative sources. Our media team places the coverage those citations come from, which is the part of GEO a technical vendor cannot supply.

A Path to the Full Program

When the diagnostic shows the gaps run wider than one discipline, the engagement steps up to The Fourth Hemisphere: Marketing for AI, where we read your whole marketing strategy against what the engines do and direct the specialists who deliver each part.

What is generative engine optimisation (GEO)?

GEO is the practice of improving how AI engines such as ChatGPT, Claude, Gemini, Copilot, Perplexity, Google AI Mode, and Google AI Overviews represent a company in the answers they generate. It combines structured, retrievable content, consistent entity data, and earned media in the sources the engines cite, so the answer a buyer receives names you and describes you accurately.

What is answer engine optimisation (AEO)?

AEO is the same discipline as GEO, named for its destination. The answer engine is whatever composes the response your buyer reads, whether a chatbot, an AI Overview, or a voice assistant. The deliverable is identical: presence and accuracy in generated answers. Treat a provider quoting GEO and AEO as separate services with caution.

How is GEO different from SEO?

SEO earns a ranked position on a results page and is measured in rankings and clicks. GEO earns a place inside a composed answer and is measured in whether the engines name you, what they say, and which sources they cite. The technical work overlaps. The earned-media work, becoming the independent source an engine quotes, is where the two part company.

Why does a PR agency deliver GEO?

AI engines weight independent, authoritative sources over anything a company publishes about itself, so the lever that moves an answer is earned coverage. Third Hemisphere has secured over 20,000 media placements since 2015, and we built the PACT diagnostic to measure the answers directly. The technical corrections are included; the citations are the point.

How long does GEO take to show results?

Factual corrections, such as fixing wrong entity data, can move answers in weeks. A program that builds presence from nothing runs 6 to 12 months, because engines refresh their retrieval sources continuously but reweight authority over months. Every program carries quarterly re-audits so you see the movement question by question.

What does a GEO engagement cost?

The PACT diagnostic is a fixed-fee engagement, and programs are scoped from its findings, so the price follows the gap rather than a rate card. Book a discovery call and we will confirm pricing for your category and markets.

Related case studies

Related reading

Related insights