SEO when search answers itself
AI answers now sit above many search results. Some things about SEO have changed. Many have not. Here is how to tell the difference.
By the Mahnovas Digital studio. Published
Search results look different from a few years ago. Google shows AI Overviews above many results, ChatGPT and Perplexity answer questions with cited sources, and many people now start research inside an AI assistant. It is fair to ask whether SEO still matters. It does, but some of the priorities have shifted.
What has changed
The biggest change is for simple informational questions. If someone asks for a definition or a quick fact, an AI answer often satisfies them without a click. Content that existed only to capture those clicks is losing traffic, and that trend is unlikely to reverse.
The second change is that being cited matters as well as being ranked. AI systems summarise and link to sources they consider clear and trustworthy. Appearing as one of those sources can bring fewer but better informed visitors.
What has not changed
AI search features still depend on crawling and indexing the web. Pages that search engines cannot reach, render or understand are no more visible to an AI answer than to a classic result. The fundamentals of technical SEO matter as much as ever:
- Pages must be crawlable, indexable and fast.
- Each page should have a clear topic, a descriptive title and a logical heading structure.
- Structured data should describe your organisation, services, products and articles accurately.
- Internal links should connect related pages with descriptive anchor text.
Commercial searches have also changed less than informational ones. When someone is ready to compare suppliers or buy, they still want to visit websites, read details and judge for themselves.
Practical steps
- Answer specific questions directly. Put a clear, short answer near the top of the relevant page, then support it with detail. This helps readers and makes your page easy to cite.
- Show first-hand knowledge. Original examples, your own data, real processes and named expertise are hard to summarise away and easy to trust.
- Keep your facts consistent everywhere. Your services, locations, prices and contact details should match across your site, profiles and directories.
- Check your robots.txt deliberately. Blocking AI crawlers is a business decision with trade-offs. For example, OpenAI documents that OAI-SearchBot surfaces sites in ChatGPT search, while GPTBot relates to model training. Google documents that its Google-Extended token controls use in Gemini and does not affect inclusion or ranking in Google Search.
- Measure referrals from AI tools. Many analytics tools can show visits referred by AI assistants. Track them alongside organic search, and look at what those visitors do next.
What we would avoid
Mass-produced articles written to target every possible keyword are a poor bet. They add little that an AI answer cannot already provide, and search engines have become better at recognising them. A smaller number of genuinely useful pages, each mapped to a real buyer question, usually performs better and lasts longer.
A simple test for any page: if an AI summarised it in two sentences, would a reader still have a reason to click through? If not, the page needs something only you can provide.
SEO is moving from "rank for as many keywords as possible" to "be the clearest, most trusted answer for the questions that matter to your buyers". That is good news for businesses with real expertise. If you want help mapping those questions to your pages, our SEO service starts with exactly that.
