You're Not Losing Rankings. You're Losing Visibility Elsewhere.

Microsoft Clarity's new AI Citations report offers an early glimpse into how AI systems discover and reference content, and the findings should make marketers rethink how they measure visibility. 

Microsoft has quietly launched one of the most interesting developments in search this year. 

Not another AI assistant. 

Not another AI Overview. 

Not another prediction that search is dead. 

Actual data! 

With the introduction of AI Citations in Microsoft Clarity, website owners can now see when their content is referenced by AI-powered experiences across Microsoft's ecosystem, including Copilot. 

The feature is still in its early stages, but it offers something the industry has largely lacked over the past two years: visibility into how AI systems surface and reference content. 

And the early signals suggest something important. 

AI visibility doesn't always look like search visibility. 

Search and AI are becoming different systems 

For years, SEO has been built around a relatively straightforward model: 

• Search demand 

• Rankings 

• Clicks 

• Traffic 

The assumption has been that if you're visible in search results, you're visible online. 

AI challenges that assumption. 

Large language models don't retrieve information in the same way traditional search engines do. Rather than matching content directly against a keyword, they often gather and synthesise information from a range of related concepts and contextual signals before generating a response. 

One way this appears in Clarity is through what Microsoft calls grounding queries. 

These are retrieval queries generated by the AI system itself to help construct an answer. They're not necessarily the same as the user's original question. 

For example, a user might ask: 

"Who's the best SEO agency for Salesforce Commerce Cloud?" 

The AI system may retrieve information through a much broader set of concepts relating to enterprise ecommerce, Salesforce implementation, SEO capability, platform limitations, agency expertise and technical performance. 

The retrieval process becomes less about matching a phrase and more about building understanding. 

That's a significant shift. 

What happens when you compare AI visibility with SEO data? 

To better understand what Clarity's AI Citations report is showing, we compared grounding query data against: 

• Google Search Console 

• Ahrefs keyword datasets 

The goal wasn't to compare rankings. 

It was to compare visibility models. 

What we found was that the relationship between the datasets was far from straightforward. 

Search Console often showed stronger semantic alignment with grounding queries, particularly at impression level. Ahrefs showed far less overlap. 

More interestingly, a noticeable proportion of grounding queries appeared to sit outside both datasets entirely. 

No obvious keyword footprint. 

No ranking trail. 

Yet those queries still generated citations. 

The implication is that AI systems may be identifying and retrieving content through signals that aren't fully captured by traditional keyword-based SEO tools. 

The real shift isn't AI search 

Much of the conversation around AI has focused on whether it will replace search. 

That framing misses the bigger change. 

What we're seeing is the emergence of a parallel visibility layer. 

Traditional search is largely driven by: 

• Demand signals 

• Ranking systems 

• Explicit intent 

AI retrieval appears to place greater emphasis on: 

• Contextual completeness 

• Topical relationships 

• Answer construction 

Both systems are trying to surface relevant information. 

They're simply doing it in different ways. 

Sometimes those systems overlap. 

Sometimes they don't. 

Why this matters for marketers 

If content can be surfaced, referenced and cited without appearing in traditional keyword datasets, then rankings alone are no longer a complete measure of visibility. 

That doesn't mean SEO is becoming less important. 

It means visibility is becoming broader than SEO. 

The question is no longer just: 

• What do we rank for? 

• How much traffic did we generate? 

It's increasingly: 

• Can our content be used to answer questions? 

• Do we demonstrate expertise beyond individual keywords? 

• Is our content comprehensive enough to be referenced when AI systems construct responses? 

In other words, ranking is becoming one visibility signal among many. 

The rise of untracked authority 

Perhaps the most interesting takeaway from the Clarity data is the visibility we can't easily measure. 

Content that shows little evidence of performance through rankings or keyword datasets may still be influencing AI-generated answers. 

That creates a new category of visibility. 

Untracked authority. 

A brand can be referenced, cited and surfaced without generating the traditional signals marketers have relied on for years. 

Most measurement frameworks aren't built for that yet. 

But they will need to be. 

What should brands do now? 

The answer isn't to abandon SEO in favour of "AI SEO". 

It's to recognise that visibility now has two distinct responsibilities. 

The first remains unchanged: 

Demand capture 

• Ranking for relevant searches 

• Winning traffic 

• Capturing measurable intent 

The second is becoming increasingly important: 

Retrieval presence 

• Building topical depth 

• Creating comprehensive content 

• Demonstrating expertise across related concepts 

• Making information easy for AI systems to understand and reference 

The brands that succeed over the next few years will likely be those that invest in both. 

Because search isn't disappearing. 

But visibility is no longer confined to search results. 

It's increasingly happening inside systems that generate answers.