Optimising For AI Overviews: Here’s What I’ve Found So Far

AI Overviews are showing up more often in Google search — and they’re starting to shift how we think about SEO performance. I’ve been tracking their growth, the types of content they pull in, and what might influence your chances of being featured.

📈 Some Numbers (March 2025 — via Semrush data)

• AI Overviews appeared in 13.14% of Google desktop searches in March — double what we saw in January (6.49%).

• Most are informational queries — 88.1% of AI Overview triggers are people looking for answers, not brands.

• Biggest growth by sector: Science (+22.3%), Health (+20.3%), People & Society (+18.8%), and Law & Government (+15.2%).

• Navigational queries are growing too, from 0.74% to 1.43%.

  
  

  Key Areas to Focus On

  
  

  1. Rethink Your Metrics

  
  

  Start tracking:

  • Keywords that trigger AI Overviews (you can spot these in Semrush, AWR, or Google Search Console).

  • How your content is displayed in the SERP (pixel width, position, features).

  • AI-triggered queries across mobile vs desktop (Overviews appear more often on mobile).

    
    

  2. Match Intent & Anticipate Questions

  
  

  AI Overviews are built to answer the question people ask, and the ones they haven’t yet asked.

• 
  

  • Create content that tackles the main question, but also naturally answers follow-ups.

  • Think about “what comes next” — e.g. after “symptoms of X”, users might ask “should I see a GP?”

  • Use multiple URLs to target different but related intents — you might get several citations in a single result.

    
    

  3. Structure is Key

  
  

  How your content is laid out matters more than ever.

  For example:

  • A clear heading structure (H1, H2, H3)

  • Short, focused paragraphs (one point per paragraph)

  • Lists, bullet points, and tables

  • Q&A or how-to formats

  • Summary or definition-style intros at the top

  • Semantic cues like “Step 1”, “Key takeaway”, “In summary”

    
    

  Avoid clutter. Pop-ups, CTAs, and anything that gets in the way of the core content can reduce your chances of being pulled in.

  
  

  4. Cover the Topic Properly (and Cut the Fluff)

  
  

  AI Overviews like depth and clarity — not just surface-level answers. But that doesn’t mean waffling.

• 
  

  • Cover the topic fully, including related subtopics and different search intents.

  • Get to the point quickly. LLMs prefer clear, direct answers over rambling intros or over-explaining the obvious.

  • Use short sentences, simple language, and ditch the filler.

  • Support key ideas with helpful extras — blog posts, how-to guides, reviews, or knowledge base articles.

  • Make sure each paragraph is doing a job - no space-fillers.

    
    

  Read about information gain here

  
  

  5. Build Authority and Trust

  
  

  AI tools don’t just pull from your site — they pull from across the web.

  
  

  • Get mentioned on trusted, relevant sites.

  • Build backlinks from high-authority domains.

  • Strengthen your digital footprint: Wikipedia pages, review platforms, YouTube, digital PR etc.

    
    

  6. Add Structured Data

  
  

  While LLMs don’t need schema to read your page, it still helps with visibility and SERP features.

  
  

  • Use FAQ, article, and breadcrumb schema where relevant.

  • Make sure your technical SEO basics are solid, especially mobile optimisation. A high share of AI Overviews come from mobile queries.

    
    

  7. Watch and Adapt

  
  

  This space is moving fast. AI Overviews are still early, and things are changing month to month.

  
  

  • Keep an eye on what triggers them in your space.

  • Test different formats, layouts, and angles.

  • Track your performance in tools like AWR (Advanced Web Ranking) and Semrush — and start building a view of your AI Overview footprint.