Technical guidance for digital teams

Generative Engine Optimisation, in brief (for government)

Practical direction for government digital and communications teams to ensure public information stays visible, cited, and trusted within AI-powered search experiences.

Last updated: November 2025

About this guidance

This guidance helps government digital and communications teams optimise content (make it friendly) for AI-powered search systems. It covers technical requirements, content structure, and implementation steps. There is a detailed version of this document available for deeper understanding.

You should use this guidance if you:

  • manage government websites or digital content
  • work in digital communications
  • are responsible for ensuring public information is discoverable

Why this matters

AI-powered search engines now answer user queries directly without requiring clicks to websites. Research shows:

  • 60% of searches end without any click
  • AI Overviews reduce click-through rates by 34.5%
  • Over 50% of searches are expected to use AI systems by 2026

Government content must remain accessible through these new channels to serve the public and internal audience (e.g. using RAG'ed) effectively.

What is generative engine optimisation

Generative engine optimisation (GEO) is the practice of structuring content so AI systems can:

  • find and understand it
  • cite it accurately in responses
  • recommend it to users

Unlike traditional search optimisation, GEO focuses on being cited within AI-generated answers rather than ranking in search results.

Three key objectives

  1. Citation inclusion: ensure your content is referenced when AI answers relevant queries
  2. Answer prominence: position content as an authoritative source
  3. Entity recognition: establish your organisation as a trusted source

Technical requirements

Use structured data

You must implement structured data markup using Schema.org vocabulary. This helps AI systems understand your content.

Priority schema types:

  • Article: for news and informational content
  • FAQPage: for questions and answers
  • HowTo: for step-by-step instructions
  • Organization: for entity information

Use JSON-LD format for structured data. This separates markup from HTML and reduces errors.

Implementing Article schema

{
  "@context": "https://schema.org",
  "@type": "Article",
  "headline": "Title of article",
  "author": {
    "@type": "Organization",
    "name": "Cabinet Office"
  },
  "datePublished": "2025-01-15",
  "dateModified": "2025-01-15",
  "publisher": {
    "@type": "Organization",
    "name": "Cabinet Office"
  }
}
          

Implementing FAQPage schema

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [{
    "@type": "Question",
    "name": "Question text",
    "acceptedAnswer": {
      "@type": "Answer",
      "text": "Answer text"
    }
  }]
}
          

Validate your markup

You must validate structured data before publishing. Use:

  • Google Rich Results Test
  • Schema Markup Validator

Common errors to avoid:

  • missing required properties
  • incorrect data types
  • malformed JSON syntax

Maintain technical accessibility

Your content must:

  • use HTTPS
  • be mobile responsive
  • meet Web Content Accessibility Guidelines (WCAG) 2.1 level AA
  • load efficiently
  • use semantic HTML

These requirements ensure both traditional search visibility and AI citation eligibility.

Content structure

Use the answer-first approach

Place key information in opening sentences. Supporting detail follows.

Example structure:

  • First sentence: direct answer
  • Following sentences: supporting evidence
  • Final sentences: additional context

This enables AI systems to extract core information quickly.

Write clear headings

Headings must:

  • describe the content below
  • use question format where appropriate
  • follow proper hierarchy (H1, H2, H3)

Good:

How to apply for a passport

Poor:

Applications

Use lists and short paragraphs

Break information into:

  • bullet lists for related items
  • numbered lists for sequential steps
  • short paragraphs (2-4 sentences)

This improves scannability for both users and AI systems.

Structure for modularity

Each section should:

  • address a discrete topic
  • work independently
  • have a clear heading
  • provide complete information

This allows AI to cite specific sections accurately.

Content requirements

Plain English

You must write in plain English. This means:

  • short sentences
  • active voice
  • simple words
  • one idea per sentence

Avoid jargon. Define technical terms when necessary.

Answer user questions directly

Identify common user questions. Provide clear, complete answers.

Format questions naturally:

  • "How do I..."
  • "What is..."
  • "When should I..."

Keep information current

AI systems favour recent content. You should:

  • review content quarterly
  • update statistics and examples
  • document changes using dateModified in Article schema
  • display last updated dates prominently

Provide accurate information

You must ensure factual accuracy. AI systems may cite your content in responses to public queries.

When information is uncertain:

  • acknowledge limitations
  • explain context
  • provide sources
  • avoid absolute claims

Authority and entity recognition

Maintain consistent information

Your organisation information must be identical across:

  • your website
  • social media profiles

This includes:

  • official name
  • address
  • contact details

Use Organization schema

Implement Organization schema on your homepage:

{
  "@context": "https://schema.org",
  "@type": "GovernmentOrganization",
  "name": "Department name",
  "url": "https://www.example.gov.uk",
  "logo": "https://www.example.gov.uk/logo.png",
  "contactPoint": {
    "@type": "ContactPoint",
    "telephone": "+44-xxx-xxx-xxxx",
    "contactType": "public enquiries"
  },
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "Street address",
    "addressLocality": "City",
    "postalCode": "Postcode",
    "addressCountry": "GB"
  }
}
          

Establish expertise signals

Demonstrate authority through:

  • author credentials on content
  • links to authoritative sources
  • clear departmental attribution
  • publication of original research or data

Platform-specific considerations

Google Search Generative Experience

Google's AI Overviews draw from traditional search index. Your existing SEO foundations remain relevant.

Priorities:

  • FAQ schema for question-focused queries
  • comprehensive topic coverage
  • regular content updates
  • clear source attribution

Bing Copilot

Microsoft's Bing Copilot emphasizes:

  • recent content
  • clear author attribution
  • factual accuracy
  • explicit structure

ChatGPT and other AI assistants

These systems favour:

  • clear, unambiguous language
  • explicit definitions
  • structured sections
  • primary source information

Implementation roadmap

Part 1: Assessment

Part 1.1: technical audit

  • verify HTTPS implementation
  • confirm mobile responsiveness
  • review existing structured data
  • identify schema gaps
  • assess current content structure

Part 1.2: baseline measurement

  • identify 20-30 priority queries
  • test queries across AI platforms
  • document current citation presence
  • record results for comparison

Part 1.3: content inventory

  • audit existing content
  • assess structure against GEO requirements
  • prioritise pages for updates
  • identify content gaps

Part 2: Technical implementation

Part 2.1: structured data deployment

  • implement Article schema on content pages
  • add Organization schema to homepage
  • deploy FAQPage schema where appropriate
  • add HowTo schema to instructional content

Part 2.2: validation

  • validate all schema implementations
  • correct any errors
  • test across devices
  • document schema templates

Part 2.3: entity information cleanup

  • establish standard organization information
  • update all owned properties
  • correct any inconsistencies

Part 3: Content optimisation

Part 3.1: restructure content

  • implement answer-first format
  • add clear, descriptive headings
  • break long paragraphs into shorter sections
  • convert dense text to bullet lists
  • add summary sections

Part 3.2: FAQ development

  • create FAQ sections for priority pages
  • implement FAQ schema
  • ensure questions match user queries
  • provide complete answers

Part 3.3: review and testing

  • verify all changes
  • test content readability
  • ensure accessibility compliance
  • check mobile display

Part 4: monitoring and refinement

Part 4.1: performance tracking setup

  • establish monitoring process
  • create tracking spreadsheet
  • configure available analytics
  • set baseline metrics

Part 4.2: post-implementation review

  • re-test baseline queries
  • document improvements
  • identify remaining gaps
  • plan next optimizations

Part 4.3: ongoing planning

  • establish quarterly review schedule
  • plan content calendar
  • define success metrics
  • document lessons learned

Measuring performance

Key metrics

Track:

  • citation frequency: how often content appears in AI responses
  • citation prominence: position within AI answers
  • attribution quality: whether citations include links
  • query coverage: percentage of target queries generating citations

Monitoring methods

Use:

  • manual query testing across platforms
  • tracking spreadsheets for results
  • Google Search Console (where available)
  • regular baseline comparisons

Monitor quarterly or when publishing significant new content.

Common mistakes to avoid

Incorrect structured data

Errors prevent AI systems from understanding content.

You must:

  • validate before publishing
  • use correct property types
  • include all required fields
  • match markup to visible content

Inconsistent entity information

Variations in organization name or contact details confuse entity recognition.

Solution:

establish and maintain standard formatting across all platforms.

Poor content structure

Dense paragraphs and unclear headings hinder AI extraction.

Solution:

use modular structure with clear headings and short paragraphs.

Neglecting mobile experience

AI systems evaluate mobile usability.

Solution:

ensure responsive design and test thoroughly on mobile devices.

Outdated content

AI systems favour current information.

Solution:

review and update content regularly, documenting changes.

Legal and accessibility requirements

Accessibility

You must comply with:

  • Public Sector Bodies (Websites and Mobile Applications) Accessibility Regulations 2018
  • WCAG 2.1 level AA

This includes:

  • alternative text for images
  • proper heading hierarchy
  • sufficient colour contrast
  • keyboard navigation
  • screen reader compatibility

See GOV.UK guidance on accessibility requirements.

Data protection

When implementing structured data:

  • follow GDPR requirements
  • limit personal information
  • obtain consent where needed
  • comply with data protection regulations

Accuracy obligations

When AI systems cite your content, you bear responsibility for accuracy.

You should:

  • prioritize factual accuracy
  • acknowledge uncertainty appropriately
  • regularly review and update content
  • correct errors promptly

Getting help

Support

For technical support:

  • consult digital team
  • refer to GDS guidance on content design
  • contact the Government Digital Service for GOV.UK-related queries

For communications strategy:

  • consult Government Communication Service guidance
  • speak to your departmental communications team

Further reading

Relevant guidance:

  • Content design: planning, writing and managing content (GOV.UK)
  • Make your website or app accessible and publish an accessibility statement (GOV.UK)
  • How to write well for your audience (GOV.UK)
  • Government Digital Service style guide

Related standards

This guidance supports:

  • Government Digital Service standards
  • Government Communication Service standards
  • Web Content Accessibility Guidelines (WCAG) 2.1

Updates to this guidance

This guidance is reviewed on []. Check this page for updates.

To suggest improvements, contact mizansyed[@]gmail.com.