My thoughts on AI Mode, and the future of Search

Google’s new AI Mode has launched, and while it’s still early, one thing is already clear: this isn’t just about SEO anymore.

The strategies we’ve depended on to drive visibility, traffic, and performance are becoming obsolete.

This article offers a snapshot of the situation, with early observations, technical takeaways, and strategic insights based on what’s surfaced so far.

AI Mode isn't perfect

I’ve already discovered and reported bugs involving rel="no-referrer" links, which is now a documented issue and known in the community.

Some of what we see seems like holdover behavior from earlier testing environments. The rollout happened quickly, and parts of the experience feel incomplete. Still, this isn’t going away. AI Mode is likely here for the long haul.

Be ready for instability and constant changes.

Optimization isn’t over, it just looks different

Traffic from language models like ChatGPT is already having a real effect. In a recent study, I found that traditional organic traffic still converts better on average for key goals.

But traffic from LLMs isn’t far behind and has shown stronger than expected performance in engagement and conversions. This proves we still have ways to influence outcomes. We just need to shift how we do it.

We have no first-party data

There’s been no official information about how AI Mode is performing. No stats, no public feedback from Google, and no leadership commentary. During the beta phase of SGE, figures like Sundar Pichai provided updates on adoption and satisfaction. This time, the rollout has been silent. That leaves us guessing.

Query Fan-Out is a BIG thing

AI Mode and AI Overviews rely on a Query Fan-Out model based on Google’s patent for generative summaries.

Here’s the process.

When someone enters a query, Google generates a set of related background questions.

It gathers content from cached pages and feeds chunks into Gemini 2.5 to generate a response.

This has major implications.

We are no longer optimizing for one query or one user intent. We are targeting a web of related questions, many of which remain invisible.

Content strategy needs to adapt

This forces a completely different approach to content creation. The old strategy of one keyword per page doesn’t hold up in a multi-query environment. We need to create content that demonstrates broad semantic relevance.

That might mean building comprehensive pages that cover several perspectives, or crafting tightly connected assets that address specific angles.

Either way, we’re building for systems that interpret meaning, not just match keywords. Start by analyzing what Gemini appears to prioritize and build out relevance from that foundation.

SEO metrics of old are no longer the north stars

Like we lost keyword data years ago, we are now watching traffic-based KPIs lose clarity.

Rather than focusing only on organic traffic, start measuring:

• Direct traffic trends to priority pages
• Share of voice across new platforms
• Market and channel reach
• Growth in brand search and recognition
• Organic traffic and performance in classical Search
• Monitoring of, and awareness of how LLMs, AI Overviews, and AI Mode are responding for our brand and user paint points we look to solve
• Optimizing the key data points in which AI Mode draws its information

Clickstream data may become one of the only effective tools for tracking user behavior.

Brand is now a bigger part of the SEO conversation

Technical SEO won’t be enough. If users and algorithms don’t recognize your brand, your content likely won’t appear.

Brand-building is now essential.

You need visibility across channels, mentions in reliable sources, and some level of cultural awareness. This isn’t about link building. It’s about being top of mind.

Efforts should include:

• Digital and traditional PR
• An active social media presence
• Strategic collaborations that expand reach

If your brand isn’t part of the broader conversation, it won’t be part of the AI-generated results either.

Final thoughts

AI Mode is still developing. It’s imperfect and adoption might be slow right now. But it marks a clear direction for the future of search. The landscape is shifting, and so must our strategies.

Instead of focusing purely on traffic, invest in layered, system-recognizable relevance.

Make your brand easy to find and hard to forget. Recognize that success now depends on influencing how search systems interpret and serve information, not just on climbing rankings.

This isn’t the end of SEO. It’s the start of a new chapter.