In an exclusive interview with Semafor, Elizabeth Reid,  who leads the Search organization at Google, detailed how the company is reshaping its core product—Search by integrating large language models (LLMs) and  generative AI to deliver synthesized responses, predictive reasoning and multi-step understanding across billions of user queries.

The transformation is not theoretical. It is already rolling out through AI Overviews, Search Live, and a dedicated AI Mode, signaling a deep architectural shift in how Google processes intent, renders results and measures relevance.

But this evolution raises urgent technical and operational questions. If generative AI is now sitting at the heart of the search engine that powers over 90% of the world’s web queries, what does that mean for traditional ranking systems?

Toggle

Free SEO Audit: Uncover Hidden SEO Opportunities Before Your Competitors Do

Gain early access to a tailored SEO audit that reveals untapped SEO opportunities and gaps in your website.

Analyze My Site

Is Google Already in the AI Search Era?

Reid’s response is unequivocal: “We are in the AI search era and have been for a while.”

This might sound like news to some, but under the hood, Google has quietly embedded AI into search for years, starting with BERT, then MUM and now with AI Overviews and AI Mode.

The real difference today? AI is no longer just an enhancement but it’s the main act.

AI Overviews now appear in a growing share of queries, summarizing complex topics, providing action steps and offering context that used to require digging through multiple sites. 

According to Reid, queries that trigger AI Overviews have jumped by 10% in the US and India—a clear sign that users are responding to this new format.

How Is Google Redefining the User’s Search Experience?

This is not about longer results but it is about lowering the cognitive barrier to asking a question. Reid explains it simply: “If you make it easier to ask questions, people just ask more questions.”

Before, search was keyword-driven. Now, with AI, users are encouraged to type what’s on their mind even if it is long, multi-layered or nuanced.

For instance, instead of searching “symptoms of vitamin D deficiency,” users can now write:

“I’ve been feeling tired lately, staying indoors, and not eating well. Could this be vitamin D deficiency or something else?”

And Google, via AI Overviews, parses the whole question and provides synthesized, thoughtful answers and sometimes drawing from dozens of sources.

This format means search is moving from index matching to intent matching, something that wasn’t fully possible before.

Are Google’s AI Features a Response to Competition from ChatGPT and Perplexity?

Well let’s be honest every product manager in Silicon Valley has asked this: Is Google still dominant in a world of ChatGPT, Perplexity and Claude?

Reid is not dismissive about it. In fact, she admits that users have alternatives. But she also offers a sharp reality check: many of these tools rely on Google’s infrastructure.

“They’re piggybacking off Google,” she says. They’re not organizing the web themselves. They’re building answers on top of what Google has already indexed, sorted, and made legible.

This context matters. 

While others may generate answers faster or in more conversational formats, Google is the only platform combining deep infrastructure with AI reasoning and it is doing both at scale.

What Is AI Mode and Why Does It Matter for Everyday Users?

Here is where the change becomes tangible. Google’s new AI Mode is a dedicated environment for users who want a more generative, conversational and dynamic experience.

What makes AI Mode different?

• It allows for multi-turn dialogue
• It presents summaries alongside traditional links
• It supports voice-based interactions via the new Search Live feature

Imagine asking, “What’s the best time to visit Italy with a family, considering school schedules and mild weather?” and receiving not just a paragraph but options, suggestions, pros and cons and direct links to flight search tools.

That is AI Mode and designed for users who want the richness of conversation and the precision of research.

Is Google Abandoning Classic Search and Blue Links?

This is one of the most common user fears. And Reid addresses it head-on.

No, blue links are not going away. Not everyone wants generative AI. Some users still prefer clicking through to a direct website, comparing sources and drawing their own conclusions.

To address this, Google is running multiple search experiences in parallel:

• Classic Search for those who want reliability and predictability
• AI Overviews for those seeking curated answers
• AI Mode for early adopters hungry for exploration

It is not about “one size fits all.” It’s about meeting users where they are, whether they are browsing on a budget phone in rural India or building prompts with GPT APIs in Silicon Valley.

Can We Trust the Accuracy of AI Overviews?

This is where nuance kicks in.

AI Overviews are not perfect and Reid doesn’t pretend they are. But they are improving fast and Google has multiple layers of verification, warnings and source linking built in.

When users ask sensitive questions (e.g., medical concerns), AI Overviews will include disclaimers, urging users to consult professionals. “We try to give you an answer, but we also give you paths to dig deeper,” she says.

It is a balance: summarize the web but don’t replace it.

Reid likens it to Wikipedia’s evolution. Once dismissed as unreliable, now it is a trusted first step for millions. LLMs, she believes, are on the same path.

What Does This Mean for SEO and Content Creators?

Reid doesn’t dodge the topic of SEO.

Yes, people are already trying to optimize for AI-generated results. Some are trying to game the system. But SEO, in her view, is not inherently bad but about making content discoverable and valuable.

That said, spam and manipulation aren’t new. “There was locksmith spam in the white pages,” she jokes.

The cat-and-mouse game continues, but Google’s decades of anti-spam experience are being applied to AI search now, too.

And here’s a clear metric: Google has kept spam under 1% of results for years and plans to continue doing so in the AI era.

Will Ads Survive in an AI-Driven Search Experience?

Perhaps the most important commercial question of all: Will ads still work in AI search?

According to Reid, yes and possibly even better.

Here is why: In AI Mode, users are asking queries that are 2–3 times longer than in classic search. That gives Google much deeper insight into intent.

If a user types, “Best eco-friendly shoes under $100 for walking on city streets,” Google knows exactly what they want. That allows ads to be better matched, more relevant, and more useful.

Plus, Google is beginning to roll out ads in AI Mode, experimenting with placement, design, and performance. It is not about slapping on traditional ads. It is about integrating commercial information seamlessly into the AI experience.

Could Agentic AI Find New Monetization Models Beyond Ads?

Reid leaves this door open.

She sees promise in agentic experiences, where AI does not just answer—it helps you do. Whether that’s making a booking, scheduling a service, or planning a trip, agentic AI opens the door to premium features, partnerships and even subscriptions.

Reid frames it this way: “Build high user value first then figure out what feels like a natural way to monetize.”

Google has always been user-first. 

This won’t change. But expect new monetization paths to emerge as users engage more deeply with AI-driven journeys.

What’s the Bigger Picture for Google and the Web?

At a philosophical level, Google is no longer just organizing the web. It’s helping users reason with it. It is analyzing, synthesizing, transforming. And Liz Reid sees this moment as just the beginning.

Whether you are asking about lunch options, building a business strategy, or seeking health advice Google wants to be your trusted AI partner, not just your search engine.

And for users? The power lies in choice. You can still search the old way. Or you can let AI do more of the heavy lifting.

The web is not dying. It’s getting smarter. The only question is: are we ready to search with it? Let’s see