Help Google understand, cite and surface your content in traditional results and AI-powered search experiences.
- Google says AI features still depend on foundational SEO and helpful, reliable, people-first content.
- Clear answers, evidence and a coherent topic cluster help both readers and systems understand the page.
- No secret schema or special file guarantees inclusion in AI Overviews or AI Mode.
What do AI Overviews and AI Mode change?
AI Overviews provide a summary inside the results page, while AI Mode lets users continue with follow-up questions. These experiences can synthesise several sources and lead users to pages that support deeper exploration. Search is therefore not always one query, ten blue links and one click. A person may get an initial answer and visit a source only for evidence, examples or a specific service.
Google's official guidance nevertheless says there are no special extra requirements for AI features. A page still needs to be crawlable, indexable and compliant with Search policies. Helpful, reliable content, a sound page experience and clear internal structure remain foundational. AEO and GEO may be useful labels, but they do not replace technical SEO.
- Be crawlable and indexable before discussing AI visibility.
- Match search intent instead of stuffing fashionable phrases.
- Render the main content in accessible HTML.
Optimising for AI search is not writing more for robots. It is making evidence, structure and expertise clear enough for people and systems to verify.
Write answers that can be cited
Begin a section with a direct answer, then explain conditions, examples and exceptions. Use headings that describe the question, concise paragraphs, checklists and tables when several attributes need comparison. This structure is not a trick for AI; it helps readers scan and helps systems identify the passage that answers a query.
Trust comes from evidence. Link to primary sources for data or technical advice, show the update date, identify the reviewing team and distinguish Friday Works' interpretation from external facts. For fast-moving subjects, use a year in the title only when there is a review plan. One maintained guide is more valuable than several new pages repeating the same information.
- Place a concise definition after a “what is” heading.
- Use examples with context, limits and measurable outcomes.
- Keep references accessible and relevant to the nearby claim.
Build topic clusters, not isolated posts
A technology company rarely establishes expertise with one “what is AI” article and one “what is SEO” article. Choose clusters connected to real delivery capability. An AI-for-business cluster might include adoption, agents, RAG, data governance, prompt security and case studies. A useful hub lets readers move from concept to implementation.
Each supporting page should serve a distinct intent and connect to a service when relevant. Avoid several URLs that answer the same question because they compete with one another. Internal links need descriptive anchor text and a two-way structure: broad pages lead to detail, while detail returns to the hub or pillar.
- Pillar: defines the subject and the main paths.
- Supporting page: answers one question or use case.
- Proof: case study, process, checklist or first-party evidence.
Technical SEO and structured data
Every article needs a self-referencing canonical, correct hreflang for available languages, unique title and description, one H1, descriptive image alt text and consistent publish/update dates. A sitemap should contain only canonical indexable URLs. Search results, filters and thin variants should be noindex while keeping links followable.
Use BlogPosting, BreadcrumbList and FAQPage when that content is visible, but do not treat schema as an entry ticket to AI results. Structured data describes entities and relationships; it does not replace the page. Test server output so titles, main copy and internal links are available without a client-side interaction.
- Keep canonical, hreflang and sitemap consistent.
- Make schema reflect visible page content.
- Monitor Core Web Vitals and mobile UX with field data.
A 90-day checklist and measurement plan
During the first 30 days, audit indexability, group queries by intent and publish one sufficiently deep topic cluster. In month two, add internal links, references, case studies and updates to pages gaining impressions but weak clicks or positions. In month three, expand queries that create leads, consolidate overlapping content and improve pages with weak engagement.
Search Console remains the primary source for Google queries, pages, clicks and impressions; avoid inventing an undefined “AI traffic” KPI. Add qualified leads, conversion by landing page and the questions prospects bring into sales conversations. The goal is not one screenshot of an AI Overview. It is durable discovery and trust in subjects that contribute to revenue.
- Month 1: technical foundation and first cluster.
- Month 2: evidence, links and content refreshes.
- Month 3: lead measurement, consolidation and expansion of winning topics.
FAQ
Frequently asked questions
Do we need separate GEO or AEO in addition to SEO?
Not as a separate foundation. Strengthen technical SEO, evidence, clear answer structure and crawlability first, then measure visibility across the relevant search experiences.
Which schema gets a page into AI Overviews?
No schema guarantees inclusion. Use accurate structured data that reflects visible content so search systems can better understand the page.
Can AI-generated content rank?
Google focuses on quality and usefulness rather than the production method alone. Content still needs verification, practical value and protection against scaled low-value publishing.
References
Sources used in this guide
We prioritise official guidance and primary technical sources. Visit each source for full context and the latest updates.