GEO Standard v1.0
GeoAI Optimization GEO Standard
1. Scope
This Standard sets out requirements for planning, delivering and measuring Generative Engine Optimization (GEO) work. It is used to assess candidates for GEO Certified Professional and organisations for GEO Certified Agency.
The GeoAI Optimization GEO Standard is a proprietary professional framework developed by GeoAI Optimization. It is not an ISO standard, is not issued by a government or regulator, and is not endorsed by any AI system provider.
2. Terminology
- GEO: Generative Engine Optimization. See the glossary.
- AI system: a product that composes answers using a language model, with or without retrieval.
- Query set: the documented list of queries used for measurement.
- Mention: an answer names the organisation.
- Citation: an answer references a source document.
- Entity accuracy: the share of factual statements about the organisation that are correct.
Further terms are defined in the glossary.
3. Principles
- Measure before and after. Changes are judged against a documented baseline.
- Accuracy over presence. Being described correctly matters as much as being mentioned.
- No guarantees. Practitioners do not promise outcomes in AI systems they do not control.
- Transparency. Methods, limits and sources are disclosed to clients.
- Integrity. No deception of users, AI systems or clients. See Ethical GEO.
4. Assessment areas
The Standard assesses ten areas, grouped into four layers. The full list, with the evidence assessors look for, appears in the framework table on this page.
| Area | Layer | Assessment question | Evidence |
|---|---|---|---|
| 4.1Discoverability | Foundations | Can retrieval systems find and fetch the organisation's key pages? | Crawl access for relevant user agents; Indexable canonical pages; Sitemaps that reflect what matters |
| 4.2Entity Clarity | Foundations | Is it unambiguous who the organisation is, what it offers and how it relates to other entities? | Consistent name, description and identifiers; Organisation and product schema; Disambiguation from similarly named entities |
| 4.3Technical Accessibility | Foundations | Can the content be read without executing heavy scripts or passing interstitials? | Server-rendered primary content; Semantic HTML and headings; No content locked behind client-only rendering |
| 4.4Content Quality | Substance | Does the content answer real questions accurately, specifically and in a form that can be quoted? | Direct answers near the top; Defined terms and stated scope; Dates, authorship and update history |
| 4.5Citation Potential | Substance | Does the organisation publish information others would need to cite: data, definitions, methods? | Original data or methodology; Stable, linkable URLs per claim; Clear attribution of external sources |
| 4.6Source Authority | Signals | Do independent, credible sources describe the organisation, and do they agree with it? | Coverage in sources relevant to the field; Accurate third-party profiles; No manufactured mentions |
| 4.7Brand Consistency | Signals | Is the organisation described the same way across its site, profiles and partner pages? | Aligned descriptions across owned profiles; Current facts (locations, leadership, offerings); Resolved outdated or conflicting pages |
| 4.8Reputation Signals | Signals | What do reviews, forums and comparisons say, and is the organisation's own response visible? | Genuine reviews on relevant platforms; Public responses to criticism; No incentivised or fake reviews |
| 4.9AI Retrieval Readiness | Operations | Is content structured so passages can be retrieved and understood out of context? | Self-contained sections with clear headings; Tables and lists for comparable facts; Machine-readable policies and specifications |
| 4.10Monitoring & Measurement | Operations | Is visibility tracked against a defined query set, with methods that can be repeated? | Documented query sets and sampling; Separate tracking of mentions, citations and accuracy; Reporting that states its limits |
5. Requirements
5.1 Professionals must demonstrate knowledge of every area in section 4 through the written assessment, and accept the code of conduct.
5.2 Agencies must additionally show:
- a documented GEO delivery workflow;
- at least one GEO Certified Professional on staff (this requirement starts 12 months after the programme opens);
- client reporting that states methods and limits;
- a complaints process;
- written acceptance of the ethical requirements by leadership.
6. Assessment criteria
Agency assessments score seven criteria: Strategy, Technical, Content, Entity, Measurement, Ethics and Process. Each is rated not demonstrated, partially demonstrated or demonstrated. Certification requires demonstrated for Ethics and Measurement and at least partially demonstrated for all others, with no more than two partials.
7. Measurement framework
Measurements reported under this Standard must name the AI systems, dates, locale and conditions sampled; publish or provide the query set to the client; report repeated runs as shares; and report mention, citation, recommendation and accuracy separately. Composite scores may be shown only alongside their components and weights.
8. Ethical requirements
Certified parties must not guarantee placements in AI answers, fabricate citations, reviews, credentials or data, create deceptive entities or sources, conceal paid placement, or misreport results. The full text is in the code of conduct.
9. Documentation
Agencies keep, for each GEO engagement: the agreed query set, baseline and follow-up measurements with dates, changes made, and client reports. Records are kept for the certification period and are sampled at renewal. Client confidential information is not disclosed to GeoAI Optimization beyond what the agency chooses to redact and share.
10. Relationship to certification
Certification attests that, at the time of assessment, the holder met this Standard's requirements. It does not attest to results for any client. Certificates state the Standard version they were assessed against. Holders are reassessed against the current version at renewal.
Version history
- 1.001 October 2026First published version.
- 1.1PlannedGuidance on multilingual query sets. Draft for public comment before adoption.
- 2.0Not scheduledReserved for structural changes, following a public comment period.
Related: certification, ethical GEO, glossary, design decisions for v1.0.