SEO alone no longer decides what AI recommends

Digital visibility is gaining a new layer.

For years, the key metrics of digital visibility have been familiar ones: search rankings, organic traffic and conversions. They still matter, but a new layer of discoverability is emerging alongside them.

More and more purchases now start with an AI's answer. That makes one thing decisive: where does AI form its picture of your company and products in the first place, whether you're a retailer, brand, or manufacturer?

Search rankings no longer tell the whole story

I wrote earlier about how competition is gradually shifting from search results to AI recommendations. When customers ask ChatGPT, Gemini or Copilot what to buy, the next question is unavoidable: how do we know whether AI recommends our products in particular?

Many people test this simply by asking an AI a few questions. Does our brand show up in the answer? What about competitors? It's a good starting point, but ultimately tells you very little. Generative AI doesn't work like a traditional search engine, and its answers can't be judged on search rankings alone.

Generative AI doesn't hand the user a list of links, it forms an answer. In doing so, it judges which products, companies and sources it considers reliable enough to build that answer on. A good position in search results therefore no longer automatically means good visibility in AI answers.

Why AI finds different things than search engines

One major reason lies in how generative search works.

Google has said it uses a so-called query fan-out technique in AI Mode. The user's question isn't handled as a single search; instead the AI breaks it into more specific sub-questions and looks up answers to them simultaneously.¹ A single user question can therefore trigger dozens or even hundreds of internal searches, whose results the AI assembles into one answer.

If a user asks, for example, "what's the best cordless drill for an electrician?", the AI can simultaneously look for information on drills suited to professional use, battery compatibility, user experiences, manufacturer comparisons and independent product reviews. The final answer is built from all of these.

This explains why the sources AI uses often aren't the same as the top of Google's search results. Independent analyses confirm the same: in Ahrefs' study, only about 12% of the sources AI cites also appear in Google's top ten.² Moz's large-scale Google AI Mode analysis reaches the same conclusion: AI builds its answers from a far wider set of sources than traditional search.³ The exact percentages will surely change as the services evolve, but the direction of the phenomenon already looks clear.

Technical search engine optimization, high-quality content and structured data are still the foundation that AI, too, builds its understanding on.

AI looks for confirmation, not just facts

This is perhaps the heart of the whole shift.

When a search engine looked for pages, it was often enough that the right keywords appeared in the right place. Generative AI aims to form a complete picture. It compares information, seeks confirmation and looks for context.

In practice, this means a single good product page may no longer be enough. If the same information, use cases and experiences are found consistently across dealers, product feeds, manuals, customer reviews and other independent sources, the AI has more material from which to form a reliable overall picture of the product. So it isn't only about the amount of information, but above all about its consistency.

AI forms its picture from your whole digital footprint

Your website is still an important source, but only one part of a much larger whole. AI combines information from a company's own content, structured product data and product feeds, product information published by dealers, customer reviews, news, expert content, videos and public discussions.

HOW AI FORMS ITS PICTURE

AI doesn't evaluate your website. It evaluates your digital footprint.

Consider a tool manufacturer launching a new cordless drill. Within a few weeks the product already lives in dozens of places: on the manufacturer's own site and in its PIM system, in dealers' online stores, in Google Merchant feeds, in YouTube demos and manuals, in customer reviews and perhaps in industry media comparisons. When a customer then asks an AI which drill is best for an electrician, the AI no longer builds its picture on a single product page,  it assembles its picture from this whole distributed footprint.

That's exactly why the digital footprint is no longer marketing's responsibility alone. It's shaped at the same time by product data, content, integrations, dealers, customer experiences and the entire digital commerce ecosystem.

Sources also vary by platform. In English-language analyses, Reddit stands out in several studies as a source type used especially by Perplexity and Google's AI Overviews, while ChatGPT appears to rely relatively more on Wikipedia. At the same time, YouTube and other video content appear ever more often among the sources behind AI answers.⁴ Different AI systems, in other words, trust different sources.

What about the Nordics?

In the Nordics there is still little comparable research. Does AI form its picture of Nordic companies primarily from their own content, from industry media, from discussion forums, or perhaps from expert content published on LinkedIn? The answer probably varies by industry.

Companies would do well to start considering where their industry's conversation actually takes place. Consumer products likely emphasise different sources than B2B solutions or manufacturing. Still, it isn't worth waiting for the research to be finished: strengthening your digital footprint is worthwhile regardless of which individual sources end up carrying the most weight in the future.

A whole new software category is emerging

When a new metric appears for a business, new tools usually appear too.

Over the past year, a whole new software category has emerged whose aim is to help companies understand their visibility in generative AI answers (among others Profound, Peec AI and Palmata).⁵ The very fact that several independent players are solving the same problem tells you AI visibility is becoming a new metric for digital commerce.

The newest tools already go beyond mere measurement: they aim to tell you what you should do next as well. Tools can therefore help identify problems and prioritise areas to develop, but visibility doesn't improve on its own. It improves only when a company's product data, content and wider digital footprint are developed based on those findings. AI visibility, then, isn't built in a single tool, but by developing the whole foundation that AI's picture is formed from.

What this means for companies

It's still too early to say what the world of AI search will ultimately look like. One thing already seems clear, though: a company's digital visibility can no longer be measured by search rankings alone.

At the same time, it's worth starting to look at your own digital footprint from a new angle. A few questions to start with:

• What information about your products is available online, and where?

• Is that information consistent across sources?

• Is there enough context around the products, and are there independent sources about them too?

AI visibility isn't a separate AI project for Solteq — it's a natural continuation of the work companies have done for years on product data, search engine optimization, content management, integrations and digital commerce. We help our customers identify which areas of digital commerce shape visibility in the AI era, and where the development work is best started.

In my previous blog, I wrote that AI can only recommend what it understands. After this blog, I'd add one more line:

AI can only recommend what it can find, understand and trust.

That's where competition in the AI era is ultimately decided.

Want to talk this through?

Get in touch →

Sources:

¹ https://blog.google/products-and-platforms/products/search/google-search-ai-mode-update/

² https://ahrefs.com/blog/ai-search-overlap/

³ https://moz.com/blog/investing-in-seo-is-geo/

⁴ https://www.tryprofound.com/blog/ai-platform-citation-patterns/

⁵ https://www.contentful.com/newsroom/contentful-introduces-palmata/