Google has recently updated their documentation on how to optimise your website for generative AI features on Google Search.
In this guide, we take a look at the key takeaways and action points, and the best ways to update and optimise your website, for better visibility in AI search features on Google.
Understand the “Gist” of AI Overviews
AI Overviews are designed to help users get the “gist” of a complex topic quickly. Google’s AI synthesizes information from multiple high-quality web sources to provide a single, cohesive answer.
Google explicitly states that AI Overviews surface links to help people explore further. Sites cited in these overviews often see a “greater diversity” of traffic because they are being recommended as expert sources for specific parts of a complex query.
Focus on “Information Gain”
Google’s AI is looking for the best answer, not just the best keywords. To be included, your content should provide:
- Unique Value: Don’t just repeat what everyone else is saying. Add original data, unique case studies, or expert perspectives.
- Structured Explanations: AI is often triggered for “complex” or “multi-step” queries. Structure your content to handle “how-to” guides and “why” questions with depth.
Technical Readiness
You don’t need a special “AI tag” to show up, but you must meet the basics:
- Crawlability: If Googlebot can’t crawl it, the AI can’t read it.
- HTTPS & Performance: A great page experience (speed and mobile-friendliness) remains a core requirement for inclusion.
- Text Availability: Ensure your most important insights are in text format, not buried inside images or videos without transcripts.
Content Structure
AI models are “extractors.” You can structure your content to better server what AI is looking for:
- Concise Page Titles: You don’t have to target specific long-tail keyword, keep your page titles on topic, and easy to read.
- Summary Meta Descriptions: Use the meta description for a brief summary, rather than a generic ‘read more about topic on our blog’ description.
- Question-Based Headers: Use H2 and H3 tags that mirror natural language questions (e.g., “What are the benefits of X?”).
- Direct Summaries: Place a concise, 40-60 word answer directly under your headers. This format is highly extractable for AI summaries.
- Anchor Links: Providing anchor links (a ‘jump to’ table of contents) helps direct AI and users through your content, and also helps AI bots identify your content.
- Lists & Tables: AI loves structured data. Use bullet points for steps and tables for comparisons.
Multi-Modal Content
Google’s AI is increasingly multimodal. This means users might search by taking a photo or asking a question about a video.
Having a featured image with your content, a video, and images throughout your can help Google’s AI bring in relevant images and video for user queries. Which means you are not just coming up as a link, your images and/or videos have potential to show also.
Managing Your Presence
If you are a brand that wants to control how much of your content is used in an AI summary, you can use traditional search preview controls:
nosnippet
Prevents a snippet (and thus likely an AI summary) from being shown.
max-snippet
Limits the length of the summary.
data-nosnippet
Prevents specific sections of a page from appearing in an overview.
See more info in our guide on how to apply search preview controls.
Summary
AI Overviews may seem like a replacement for SEO – but I see them more of an addition to the search results. Google’s goal remains the same, connecting users with the most helpful, reliable and people-first content. If you focus on being the most authoritative answer in your niche, the AI will naturally find you.
Using the best-practice SEO methods for your website are in line with what AI is looking for.
I do think that AI overviews are taking traffic away from websites – but by having your website appear in the AI overviews, this increases your brand visibility, and confirms that Google has found your content useful, informative and related to the search query.
Read Google’s full documentation here.
Found this useful? Share and comment below!
