Why technical SEO still wins in an AI search world

Every few months someone tells me SEO is finished because AI now answers the question directly. Then I open Search Console for a new client and find half their important pages are not indexed. The way results are presented has changed a lot. The way Google decides which pages are worth showing has not.

What AI Overviews actually changed

AI Overviews summarise an answer at the top of some results and link to the pages they drew from. For simple factual questions that means fewer clicks. For questions where people compare options, check credentials or are ready to buy, people still click through, and the pages that get cited are pages Google already trusts.

So the question is not "SEO or AI?" It is whether your site is one of the sources Google is confident enough to rank and quote. That confidence starts with technical health.

1. Can Google crawl and index your important pages?

The Page indexing report in Search Console is where I start. Look at the pages marked "Crawled – currently not indexed" and "Discovered – currently not indexed". If your service pages or key categories are in there, nothing else you do matters until that is fixed.

Common causes I find:

• Thin or near-duplicate pages, such as location pages with only the town name changed
• Filter, tag and parameter URLs eating the crawl budget
• Important pages with almost no internal links pointing to them
• Canonical tags pointing somewhere unexpected after a redesign

2. Is the page fast and stable on a phone?

Core Web Vitals measure how quickly the main content appears (Largest Contentful Paint), how quickly the page responds to a tap (Interaction to Next Paint) and how much the layout shifts (Cumulative Layout Shift). They are part of how Google assesses page experience, and they affect whether visitors stay. Check the Core Web Vitals report in Search Console for real-user data rather than relying on a single lab test.

3. Does the structure make the answer easy to find?

Clear structure helps people skim and helps Google understand what each part of a page is about. One H1 that says what the page is. Subheadings that match the questions people ask. A direct answer near the top, then the detail. Tables and lists where they genuinely help. This is also the format most likely to be quoted in an AI answer.

4. Does structured data describe you accurately?

Structured data tells search engines what things are: your business, your services, your articles and their author, your products and their prices. It will not rank a weak page on its own, but it removes guesswork. Test it with Google's Rich Results Test and keep it truthful; markup describing content that is not on the page breaks Google's guidelines.

5. Is it obvious who is behind the content?

Google's quality guidelines put trust at the centre of E-E-A-T. Technically, that means named authors with a real profile page, Person and Organisation structured data linking to your genuine profiles elsewhere, clear contact details, and case studies with numbers you can back up. These are small build decisions that make a large difference to how credible a site looks to people and to Google.

6. Do your pages link together by topic?

Internal links are how you show Google which pages matter and how subjects connect. A service page should link to the articles that support it, and those articles should link back. On most sites I audit, the blog and the service pages barely link to each other, which wastes the authority the articles earn.

Where to start this week

1. Open the Page indexing report and list any important pages that are not indexed.
2. Check the Core Web Vitals report for pages failing on mobile.
3. Make sure every key page has one clear H1 and a direct answer in its first paragraph.
4. Add links from your three best articles to the service pages they relate to.

If you would like a second pair of eyes, this is exactly what my technical SEO audit covers, and you can see the kind of result it leads to in this e-commerce case study.