Google AI Overviews have become one of the defining features of modern search. Since their wider rollout in 2024, they've changed how users interact with informational queries — and fundamentally changed what "appearing at the top of Google" means.
This guide covers what we've learned from optimising content for AI Overviews: what actually triggers them, the content and technical factors that influence source selection, and how to monitor whether your optimisation is working.
What Triggers Google AI Overviews?
AI Overviews don't appear for every query. Based on testing and published analysis, they're most consistently triggered by:
Informational queries — "How does X work?", "What is X?", "Why does X happen?" — queries where Google determines the user wants an explanatory answer rather than a specific website.
Complex multi-step questions — Queries that require synthesising multiple pieces of information into a coherent answer, where a single page is unlikely to cover everything comprehensively.
Research and comparison queries — "What are the differences between X and Y?", "What are the options for X?" — queries where aggregating perspectives adds value.
"How to" queries — Step-by-step instructional queries reliably trigger AI Overviews with the steps laid out in the AI-generated answer.
YMYL queries with established consensus — Health, finance, and legal queries where Google has high confidence in established authoritative sources.
Notably, AI Overviews appear less frequently for:
- Navigational queries (brand name searches)
- Purely transactional queries ("buy X", "X discount code")
- Highly localised queries ("best [service] near me")
- Breaking news queries (Google prefers traditional news results for recency)
- Highly contested or opinion-based queries
Understanding the query types that trigger AI Overviews helps you focus optimisation effort on content that will actually benefit.
How Google Selects AI Overview Sources
Google has not published a formal algorithm for AI Overview source selection. Based on testing, published research from Google DeepMind, and pattern analysis across thousands of queries, these are the factors that consistently influence citation:
Direct passage-level relevance
This is the most important factor. Google AI Overviews retrieve at the passage level, not the page level. Each section of your content is evaluated independently against the query.
A page that ranks #1 in traditional search but whose relevant section is poorly structured will lose out in AI Overviews to a page that ranks #5 but directly answers the question in its opening sentences.
The implication: optimise each section as if it were an independent answer to a question. Don't bury the answer in paragraph three.
E-E-A-T signals
Experience, Expertise, Authoritativeness, and Trustworthiness are primary quality filters. Google is highly risk-averse about what it includes in AI-generated answers — particularly for health, finance, and legal content.
Strong E-E-A-T signals for AI Overviews include:
- Named author with verifiable credentials linked to a detailed author page
authorproperty in Article schema with a Person node- Consistent authorship across a defined topic area (an author who has written ten articles on technical SEO is treated as more authoritative than one who has written one)
- External citations and press coverage of your organisation
- Third-party verification of claims (citing NICE, published research, industry bodies)
Content structure
Google AI Overviews favour content that is:
- Structured with clear H2/H3 hierarchies — Each heading should represent a distinct question or subtopic
- Concise within sections — 50–150 word sections that cover one point completely
- Formatted with lists and tables where appropriate — Step lists for processes, tables for comparisons
- Written with the direct answer first — Inverted pyramid within each section
Schema markup
FAQPage and HowTo schema provide machine-readable signals that directly map to AI Overview formats:
- FAQPage schema tells Google that a section contains explicit Q&A pairs — precisely the format AI Overviews use for FAQ queries
- HowTo schema signals step-by-step instructional content — reliably cited for "how to" queries
- Article schema with
dateModifiedprovides freshness signals
Freshness
For queries about evolving topics — AI tools, algorithm updates, platform changes, regulations — AI Overviews strongly favour recently updated content. Add a visible "Last updated: [date]" near the top of your article and keep dateModified in your Article schema current.
What We've Found Actually Works
After running GEO programmes for clients across various sectors, these are the tactics with the most consistent impact on AI Overview citation rates:
1. FAQ sections with FAQPage schema
This is the single highest-impact tactic. Adding a FAQ section of 4–6 questions — directly mapping to the People Also Ask questions that appear for your target query — and implementing FAQPage JSON-LD schema creates both a visible and machine-readable answer structure.
The key is matching FAQ questions to actual user queries (not your preferred framing of the topic). Use Google's People Also Ask boxes, Answer The Public, and Search Console question-format query data.
2. Definition-first section openings
Start each H2 section with a direct, concise definition or answer (2–3 sentences) before expanding into depth. This is the passage Google extracts.
Poor approach:
"When thinking about technical SEO, it's worth considering that there are many factors involved, and the importance of each will vary depending on your specific situation. Crawlability is one such factor..."
Better approach:
"Crawlability is the ability of search engine bots to access and process your website's content. Poor crawlability directly limits how much of your site gets indexed and can cause ranking issues even for high-quality content."
The second version is extractable. The first isn't.
3. Sourced statistics
AI Overviews de-prioritise content with unsourced claims. For every statistic or data point, use explicit attribution:
"According to [Source] ([Year]), X% of [finding]."
This serves dual purposes: it increases AI Overview citation probability and builds E-E-A-T trust signals simultaneously.
4. Google-Extended access in robots.txt
This is often overlooked. Google-Extended is the crawler Google uses for AI Overviews content. If your robots.txt blocks it, your content is categorically ineligible for AI Overview citation. Check immediately:
# Correct configuration for AI Overview eligibility
User-agent: Google-Extended
Allow: /
Many sites have inadvertent restrictions via wildcard rules. Verify your robots.txt explicitly allows Google-Extended.
5. Keeping content current
For topics where Google regularly updates AI Overviews, stale content is progressively deprioritised. Build a review cycle for your most important content — even minor freshness updates (adding a current statistic, updating an example) help maintain citation rates.
Common Mistakes We See
Optimising for the page, not the passage — Producing long, comprehensive pages that bury their answers in flowing prose. AI Overviews need extractable passages — clear headings, direct answers, concise sections.
Ignoring AI crawler access — Blocking Google-Extended (often via User-agent: *) while thinking you're only blocking GPTBot. These are separate user agents; block one without blocking the other.
Thin FAQ sections — Adding a FAQ section with one-sentence answers to tick a box. AI Overviews favour complete, substantive answers. Your FAQ answers should each be 50–100 words minimum.
Inconsistent author attribution — Publishing content without named authors or with authors whose credentials aren't verifiable. For YMYL topics especially, this is a primary disqualifier.
Not monitoring AI Overview appearances — Running optimisation without tracking whether it's working. Test your target queries manually in Google regularly, and use monitoring tools for scale.
Monitoring Your AI Overview Citations
Tracking AI Overview performance is less straightforward than tracking traditional rankings, but several approaches are available:
Manual testing — The most reliable method. Test your 20–30 priority queries in Google regularly. Screenshot AI Overviews where your content is cited.
Search Console — While Search Console doesn't explicitly label AI Overview clicks, you can infer AI Overview traffic through click-through rate changes for informational queries. Pages cited in AI Overviews often see CTR changes as AI absorbs some clicks.
Third-party tools — Platforms like Semrush and Ahrefs have introduced AI Overview tracking features. Otterly.AI and Profound offer more specialised AI search monitoring.
Referral traffic patterns — Monitor organic traffic patterns for pages you're optimising. Citation in AI Overviews can generate traffic from the citation link beneath the overview.
FAQs
Can I request that Google include my site in AI Overviews? There's no direct submission process. Optimisation for inclusion is the only lever — content structure, E-E-A-T signals, schema markup, and ensuring Google-Extended can crawl your site.
Will AI Overviews reduce my organic traffic? For queries where an AI Overview appears, click-through rates to traditional organic results typically decline. Being cited within the AI Overview partially offsets this — citation links do generate clicks, and brand impressions within AI Overviews have indirect value.
How many citations does appearing in an AI Overview generate? AI Overviews typically cite 3–8 sources. Being in the first 3 citations generates the most clicks. Citation position within the AI Overview correlates loosely with how directly the passage answers the query.
Does AI Overview optimisation help with other AI platforms too? Yes significantly. The content practices that improve AI Overview citation — answer-first structure, FAQ sections, FAQPage schema, direct passages, sourced statistics — also improve performance in Perplexity, ChatGPT Search, and Copilot. It's the same underlying content quality signal.
How long does it take for AI Overview optimisation to take effect? After Google recrawls your pages (typically within days for frequently crawled sites), changes to content structure and schema can affect AI Overview citation within 2–6 weeks. E-E-A-T signals and domain authority take longer to build.
Ready to start appearing in AI Overviews for your target queries? Our GEO services cover full-funnel AI search optimisation. Start with a free audit.
