Craig Campbell.
← All posts

3 November 2026

AI search visibility in 2026: measure the exposure, then check the business result

An AI search impression is not a lead. That sounds obvious, but it is worth putting at the top of a reporting meeting before a new dashboard starts doing the talking.

Google announced dedicated Search Generative AI performance reports in Search Console on 3 June 2026. They provide a separate view of impressions within features such as AI Overviews and AI Mode, alongside information about pages, countries, devices and dates. That is a useful development: businesses have a more direct way to inspect where their pages appear. It still leaves an important question unanswered. What happened next?

Start with what the report actually measures

The launch announcement describes visibility, not a ledger of revenue generated by AI search. A page appearing in an answer, someone visiting the website and a customer buying are three different events. Keep them separate in your reporting.

Make a simple monthly record of the pages appearing, the countries involved and changes in impressions. For a UK business, looking at UK visibility separately can be more useful than celebrating a worldwide total. Note significant site changes alongside the figures so you can investigate patterns rather than relying on memory.

Follow the landing page, not just the headline number

Next, look at those pages in your analytics. Are they attracting relevant visits? Are people reading further, reaching a service page or submitting an enquiry? If the page is a shop product page, check sales rather than treating a visit as the finish line.

This comparison is a diagnostic exercise, not perfect attribution. Search Console and analytics describe different parts of the journey. You should not assume every change in organic traffic was caused by an AI feature, or that every direct visit came from an AI tool.

Do not turn missing data into a conclusion

Reporting availability and thresholds matter. Check Google’s current help documentation before treating the absence of an AI report as proof that your site never appears. Likewise, a small number of impressions may not be enough to support a confident judgement about a content change.

Other AI services also have their own referral patterns. Where an identifiable referring source is available, inspect it separately. Where it is not, label the uncertainty. Inventing a neat explanation does not improve measurement.

The SEO work is still recognisable

Google’s guidance says there are no additional technical requirements or special optimisations needed for AI Overviews and AI Mode. The usual SEO foundations remain relevant. Before buying a special AI visibility package, check that your important pages can be crawled, explain the subject clearly and give readers a reason to trust the information.

For a consultancy page, that might mean describing the actual service, its boundaries and how to enquire. For a product page, it means reliable product information. Adding vague claims about being an industry leader does neither job.

A reporting routine worth keeping

  • Record visibility by page and relevant country.
  • Compare landing-page visits and meaningful business actions.
  • Annotate content, technical and campaign changes.
  • Investigate patterns over time rather than one isolated spike.
  • Keep visibility, attributed outcomes and proven causation separate.

The useful change in 2026 is better access to visibility information. Use it to ask sharper questions, not to manufacture certainty. If your search strategy needs a clearer connection with the business, explore SEO consultancy with Craig.

Sources and further reading