What’s actually changing with search
For the last twenty years, search has been a list. You type a query into Google, you get ten blue links, and you click the one that looks most useful. That model is finally being disrupted, not by a competitor to Google, but by Google itself, alongside a wave of AI-first search tools.
The shift is simple to describe but huge in its implications. Instead of returning a list of pages, AI search tools read multiple sources and write you an answer. Sometimes they cite the sources, sometimes they don’t. Either way, the result is fewer clicks to websites and more answers happening inside the search interface.
If you’ve not yet seen Google’s AI Overviews, try searching for something practical like “how do I fix a slow WordPress site” or “what’s the best way to track conversions.” You’ll often see a generated summary at the top before any traditional results.
The tools your customers are starting to use
- Google AI Overviews: generated answers at the top of standard Google results, rolling out across the UK throughout 2025 and 2026.
- ChatGPT search. OpenAI’s search-enabled mode that browses the web and returns cited answers.
- Perplexity. An AI-first search engine that’s grown rapidly with users who want direct answers.
- Claude: anthropic’s assistant, used heavily for research and increasingly for finding suppliers and services.
- Microsoft Copilot: built into Bing, Windows and Edge; quietly significant volume.
None of these will replace Google overnight. But collectively, they’re already redirecting a chunk of search behaviour, and the trend is one-way.
What still works (and what’s changed)
Here’s the good news: most of what we already know about good SEO still applies. AI tools don’t ignore Google’s index; in many cases they’re built on top of it or use the same crawl-and-rank logic to decide which pages to read. Strong technical SEO, clear site architecture, and authoritative backlinks all still matter.
What’s changed is how your content gets read once it’s been found.
You’re no longer writing only for a human reader who scans, scrolls and decides. You’re also writing for an AI that’s trying to extract a clean, accurate answer in two sentences.
That subtle shift is what AI search optimisation is really about. We covered the broader principles in our post on modern SEO beyond keyword stuffing, and most of those principles map directly onto AI-friendly content too.
How AI tools choose what to quote
From the research that’s emerged so far, AI search tools tend to favour content that:
- Answers questions directly. If the question is “what is domain authority”, the page should literally answer that within the first paragraph or two, not bury it under 500 words of preamble.
- Uses clear structure. Headings, short paragraphs, bullet points and definition-style sentences are easier for an AI to extract.
- Demonstrates expertise. First-hand experience, real numbers, named tools and named people all signal that the content is grounded rather than recycled.
- Cites sources. Pages that link to credible references tend to get cited more themselves.
- Has fresh dates. Recently updated content is preferred for any topic where the answer might have changed.
If you’re already writing helpful, well-structured blog posts, you’re most of the way there. If you’re churning out 300-word posts written for keyword density, AI search is going to skip past you completely. We have turned these principles into a practical SEO and AI content checklist you can run through for every page.
Practical steps to take now
1. Audit your most important pages for clarity
Pick your top five pages by traffic (or by commercial importance). For each one, ask: “If I was an AI trying to answer the question this page targets, would I find a clear, quotable sentence in the first 200 words?” If the answer is no, rewrite the opening so it does.
2. Add a TLDR or summary block to long posts
You’ll have noticed we’ve been doing this on the Notus blog for a while. A short summary at the top of an article gives both human readers and AI tools a clean overview of what the page covers. It also forces you to be clear about the point you’re making.
3. Use real data, examples and named tools
Generic advice gets ignored. Specific advice gets cited. Saying “we increased a client’s conversion rate by 32% by changing their CTA copy” is far more useful than “good CTAs improve conversions.” If you’ve got real numbers, share them. Our post on marginal gains is a good example of how small specifics add up.
4. Build out your topic clusters
AI tools tend to surface sites that demonstrate depth on a topic. One blog post about CRO is fine. Twelve interconnected posts that cover everything from the basics of CRO through to heatmaps and session recordings is a much stronger signal of authority. We talk about this approach in more detail in our guide on how keyword research transforms strategy.
5. Tighten your schema markup
Structured data has always helped Google. It’s becoming even more important for AI search tools, because it gives them an unambiguous way to understand what your page is about, who the author is, and what kind of business you run. If you’re not using schema yet, it’s worth getting set up. We’ll be covering this in more depth in an upcoming post.
6. Make sure your business is properly described elsewhere
AI tools cross-reference. They look at your website, your Google Business Profile, your LinkedIn, your local citations, and any press or directory listings they can find. If those sources contradict each other (different addresses, different services, different phone numbers), AI tools will either pick the wrong answer or leave you out altogether. Our post on local citations covers how to clean this up.
What about traffic?
This is the big question, and the honest answer is: nobody knows yet. Some studies suggest AI Overviews are reducing click-through rates on informational queries by 30 to 60 per cent. Others suggest the impact is smaller and that commercial queries (the ones that actually generate leads) are barely affected.
What seems clear is that the value of being cited is shifting. Even when an AI doesn’t send a click, being mentioned by name in an answer carries weight. People remember the brand they read about. They search for it later. They mention it to colleagues. In that sense, AI search is starting to feel a little more like brand-building than direct response marketing.
What we’d recommend right now
If you’re a small business, here’s the short version of what to do:
- Don’t panic. Most of your existing SEO work is still valuable.
- Make your content clearer, shorter where possible, and more direct.
- Add summaries and structure to longer articles.
- Keep publishing genuinely useful content on a consistent schedule.
- Watch your analytics for changes in traffic patterns and search queries.
- Test the AI tools yourself. Ask them about your industry. See what they say. See who they cite.
If you’d like a hand thinking about how AI search affects your specific business, or you’d just like a sanity check on your current content strategy, get in touch. We’re happy to talk through it.
New to some of the language around search? Our glossary of plain English SEO terms breaks down the words you keep hearing.
