Google’s head of Search reveals how artificial intelligence is fundamentally reshaping how billions discover information—expanding the pie rather than replacing traditional search, while forcing a reckoning for shallow content creators.
Executive Summary
In an extensive interview with The Economic Times, Liz Reid, Google’s VP and head of Search, laid out the company’s most comprehensive vision yet for the AI era of search. The message is clear: AI isn’t killing traditional search—it’s supercharging it. But the implications for content creators, SEO strategies, and user behavior are nothing short of revolutionary.
The Core Philosophy: Augmentation, Not Replacement
Reid delivered Google’s defining statement on AI and search at the 16-minute mark of the interview: “We don’t view AI as replacing search in the search experience. We view it as augmenting, as enabling us to reinvent search.”
I think we’re still really at the beginning. I think there’s still a huge way to think about how we use AI to reinvent search. The goal [is] making it possible that you can ask truly anything on search and that you can do so effortlessly.”
— Liz Reid, VP & Head of Google Search
This isn’t just corporate messaging. Reid emphasized that Google has been integrating AI into search for years through technologies like BERT and MUM, but the evolution to AI Overviews and AI Mode has made these capabilities more visible to users at the feature level.
The Numbers: Explosive Growth and Behavioral Shifts
User Adoption at Scale
AI Mode has already surpassed 100 million active users per month across just the United States and India within months of launch. To put this in perspective:
- US rollout: May 2025
- India rollout: Late June 2025
- Time to 100M users: Approximately 2-3 months
- Countries covered: Just 2 (with more rolling out)
Reid expressed satisfaction with the reception so far, noting “we’re only a few months out if you think about it”.
The Query Revolution: 2-3x Longer Searches
User queries in AI Mode are 2 to 3 times longer than traditional searches, but not because users feel obligated to type more—they’re providing more context because they can.
Example from Reid:
Instead of searching “restaurants” or even “kid-friendly restaurants,” users now ask: “I have a four-year-old and a seven-year-old. I need to ensure it’s family-friendly, but I need to be outside. Can you please find me restaurants?”
| Traditional Search | AI-Enhanced Search |
|---|---|
| “kid friendly restaurants” | “I have a 4yo and 7yo. Need family-friendly, outdoor seating. Find restaurants?” |
| Single keyword query | Multi-constraint contextual query |
| One search, move on | Follow-up questions encouraged |
| 3-5 words average | 9-15+ words average |
The Psychology of Search: Why People Ask More Questions
Reid revealed a fascinating insight about user behavior: “There’s this sort of idea that many people have that the number of questions you have is sort of fixed and that you’re already bringing them all. But that’s not actually what we see.”
The Implicit Calculation
Users make an implicit calculation: “Is it worth my time? Am I confident that a response is going to come? Is it going to take too much effort?” When AI lowers the bar to ask questions and get good responses, people ask more questions—including questions they wouldn’t have asked otherwise.
Reid’s Personal Example:
Reid shared how she used AI Overviews to learn about cricket because friends and coworkers would make statements about the sport and she didn’t want to reveal how limited her cricket knowledge was. She would simply type questions and get responses, enabling casual learning moments that she wouldn’t have pursued through traditional search.
“I used AI Overviews to learn a bunch about cricket. I still don’t know very much about cricket, so don’t quiz me on my cricket. But I learned various facts over the time because I had friends and co-workers who were big into cricket and they would make these statements. I’d be like, ‘I don’t know what you’re talking about.’ And so I would just type this question because I didn’t want to show up just how dumb my cricket knowledge was.”
— Liz Reid
The Follow-Up Effect
In AI Mode, users can ask follow-up questions without repeating context: “If I had to go repeat my entire context, all everything I said before, it wouldn’t be worth my time. But if I can just ask a quick follow-up, then I’ll go ask it”.
The result: People are asking more questions overall—not just harder questions.
The Blue Link Isn’t Dead—It’s Evolving
Reid was emphatic: “I do think the story of the blue link is far from over”. But the reasons people click are fundamentally changing.
Why Users Still Want Links
Reid identified several key drivers for continued link clicks:
- Human connection: “Many, many people want to hear from other people… people want to connect with humans and that human spirit”
- High-stakes decisions: For fashion advice or major purchases, users want to hear directly from creators, not just what the AI model says
- Diverse perspectives: “One answer isn’t the same for everyone”
- Direct voices: Users don’t want interpretations—they want to connect directly to creators
“I like the value of AI responses, but I also don’t like when they’re like, ‘So, you know, I’m going to tell you what so and so said.’ I’m like, ‘I want to hear what so and so said directly, right? Don’t tell me what so and so said. Connect me to that.'”
— Liz Reid
What Content Survives
Reid observed an evolution in user interests: “They want that depth of experience. They want that personal view. So they’re more interested in web content that is rich, that is not sort of like just the shallow basic fact, but brings in the unique perspective or that creator view”.
She emphasized that web content, short-form video, and user-created content will continue to flourish, and Google is experimenting with inline links within AI Overviews and AI Mode that link directly to creators.
The Traffic Quality Revolution: Bounce Clicks Are Dying
Here’s a metric that didn’t make headlines but should: bounce clicks are declining dramatically.
Reid explained: “There’s less of what we call bounce clicks, where a user clicks and they immediately click back, right? They’re like, ‘Oh, I thought this was going to be useful. No, it’s not.’ And because we’re giving enough context, people kind of know and they don’t have to do as many of those bounce clicks”.
What This Means
| Metric | Old Search | AI-Enhanced Search |
|---|---|---|
| Total clicks | Higher volume | Slightly lower volume |
| Bounce rate | High (quick back-clicks) | Lower (informed clicks) |
| Deep engagement | Mixed quality | Higher quality |
| User intent | Exploratory/uncertain | Informed/purposeful |
The implication: “What you do see is a lot of excitement on the deep clicks”.
Publisher Impact:
Reid advised publishers to shift their thinking: “As a publisher, you want to really think about like what is the user’s experience when they get to my site. Be optimizing for a great experience when you land on that page, because people are going to want to go there not to just get that 5-second thing, but to actually go deeper”.
The SEO Apocalypse: Shallow Content Is Dead
Reid delivered one of the most direct warnings to content creators in Google’s history.
The Death Sentence for Thin Content
Reid’s message was unambiguous: “If you produce content that’s very shallow, that you’re just going to hope is ranking at the top but doesn’t really have much to say, then your content really doesn’t have much more than like an AI Overview would give in the first place. So, people aren’t going to want to click to go check the same fact”.
“People should really produce content that users care about and not think about building content for search engines. But that only becomes more pronounced now. If you produce content that’s very shallow that you’re just going to hope is ranking at the top but doesn’t really have much to say, then your content really doesn’t have much more than like an AI overview would give in the first place.”
— Liz Reid
The New Content Hierarchy
Content That Wins:
- ✅ Rich, detailed information with unique perspectives
- ✅ Personal experience and expertise
- ✅ Content worth reading for more than 5 seconds
- ✅ Deep dives that go beyond surface facts
- ✅ Creator voices and authentic human connection
Content That Loses:
- ❌ Shallow, manufactured content
- ❌ Basic facts repackaged for SEO
- ❌ Content built for search engines, not users
- ❌ Quick-answer material with no depth
- ❌ Generic information available everywhere
The Guidance Evolution
Reid noted that while some guidance remains the same—”people should really produce content that users care about and not think about building content for search engines”—it’s now “more pronounced.
The shift is toward “more push on niche content, more push on the detailed understanding of the space that you’re coming in. So less manufacturing of content and more like actually putting yourself in the space”.
Trust and Provenance: Google’s Competitive Moat
One of the most revealing insights came when Reid discussed competition from standalone LLMs like ChatGPT.
The Double-Check Phenomenon
Reid revealed a fascinating user behavior: “Sometimes people don’t seek provenance on the LLMs. They come to Google and they seek their provenance. So we see a lot of people go ask the question and then they come to Google to double check it”.
What This Reveals:
- Users recognize that standalone LLMs lack source transparency
- Google’s link-based approach provides verification capability
- Trust remains a differentiator in the AI era
The Trust Imperative
Reid identified trust as non-negotiable: “We want to do everything we can to maintain that trust and to nurture that trust. So if we feel like we have to make a trade-off that breaks the trust, then we don’t want to do that”.
She also emphasized the “web forward ecosystem” as essential: “How do we bring those voices forward and allow you to get started on Google but to make it easy for you to click out to the web to go deeper throughout.
Deep Research: How Ranking Actually Works
Reid provided rare transparency into how Google’s AI-powered “Deep Research” feature works behind the scenes.
When asked if ranking works the same way as traditional blue links, Reid confirmed: “Mostly yes. What we do with something like deep search is we are querying web search underlying ranking. And so in that sense it’s the same”.
The Process
The difference: “It’s different in the sense that a question a user is asking with deep research is often more complex than any single web search will go. And so we’ll take that search query and we’ll break it into several queries”.
Deep Research Workflow:
- User asks complex question
- AI breaks it into multiple sub-queries
- Each sub-query hits core web ranking (same as traditional search)
- AI synthesizes results from multiple ranked sources
- User receives comprehensive answer with links
Reid emphasized: “We are really building on top of that long tradition and expertise and trust in web ranking, but doing it to answer questions that are not possible as a single query.
The Expansionary Moment: Everyone Can Grow
Reid pushed back hard against the zero-sum narrative.
She stated: “We’re really in this expansionary moment. The number of questions that people ask is not fixed. So there’s no sort of zero-sum nature here. Actually, as the tools get more useful, people are just asking many, many more questions, and so it’s actually possible that everyone can grow at the same time—maybe not everyone, but lots of people including Google can grow in this time”.
The AI Overview Traffic Story
Reid provided specific data: “We see with AI Overviews people issue more queries and so at the highest level the two roughly balance out” when asked about ad revenue being the same with or without AI Overviews.
The Math:
- More queries overall = more total opportunities
- Ads appear lower on page with AI Overviews
- Volume increase offsets position changes
- Net result: Revenue stability
Agentic AI: The Control vs. Automation Balance
Reid offered insights into Google’s approach to agentic features—AI that takes actions on behalf of users.
The Philosophy
Reid explained: “The way that we think about [agentic AI] in search is if you go back to this idea that you can ask for anything effortlessly, where is their grunt work? Where is there a bunch of work that a user just doesn’t want to spend time—it’s too much effort. And how do we take that work out of it?”
The Control Challenge
The key consideration: “The tech is nascent and also user expectations are changing and early on, and so we have to figure out how we both take out the grunt work but don’t eliminate control where a user wants it”.
Reid’s Examples:
For important transactions: “Maybe we can do a lot of the work to get to the point where like most of the initial form filling is there, but you maybe still want to click yes on the transaction rather than have us do the final transaction for you, right?”
For routine tasks: “Maybe something is repeat and you’re just like, ‘I just want to do this again and again.’ Fine.”
Multi-Modal Approach
Reid revealed Google’s multi-pronged strategy: “From a tech perspective, there’s not a single solution. Maybe we have to automate some of your phone calls. Maybe we have to automate some of your browser actions. Maybe there’s times where we integrate on the API side. And we should pull them all together”.
Example: Google has been working on automating phone calls for services like booking haircuts or finding plumbers, where in the US “you actually have to pick up the phone and call a whole bunch of plumbers to get out basic information.
India: From Afterthought to Innovation Leader
Reid’s comments on India reveal a dramatic strategic shift at Google.
The Timeline Transformation
| Timeframe | Launch Gap (US to India) |
|---|---|
| Few years ago | 12-18 months later |
| Now | 2 months or weeks |
| AI Mode | 1 month after US |
| Future goal | Sometimes first |
Reid stated: “If I looked a few years ago, if we launched something in the US, like I don’t know, maybe a year or 18 months later, it would come [to India]. Now, we’re really trying to shrink it down—two months, if not weeks”.
India as Innovation Hub
Reid’s vision: “We’re going to continue to push both bringing innovations rapidly to India but also thinking about where does India actually prove a ground for innovating itself—like it’s not always going to be second, you know, oftentimes it will be second and not third, fourth, or fifth, but sometimes I think it’ll even be first going forward”.
Why India Matters:
Reid highlighted India’s unique position:
- “One of our top markets overall”
- “Our biggest market for Lens”
- “Our biggest market for voice”
- Users are “really rapid at adapting a bunch of the AI innovations”
- “Users are really challenging the systems and so we really need to respond and innovate”
The Language Revolution
Google launched “one of our first cross-language experiences with AI Overviews toggling between Hindi and English” and has rolled out AI Overviews in languages beyond Hindi.
Reid explained the breakthrough: “Historically, if the content didn’t exist on the web in your language, it was kind of locked out from those users. And LLMs are allowing us to take the learnings from one language and output them in a way that’s understandable for the users”.
The Literacy Revolution: Beyond Language Barriers
Reid articulated a vision that goes beyond just translation.
She referenced Google’s founding mission: “Larry and Sergey set this vision of make information universally accessible. But it hasn’t really been universally accessible in many ways with language, but it’s also not just even across languages but also even within a language—what’s your level of literacy on something?”
Reid’s Personal Example
She shared: “I may consider myself to have fairly good tech literacy. I have okay financial literacy. I have pretty poor medical literacy, right? Like you give me a medical paper and I’m like, ‘Oh my gosh, I’m going to have to spend five hours.’ But if I can learn about a topic that matters to my daughter, that’s really powerful to me”.
The Vision: “It’s not enough for access to information if you don’t understand it, and now we’ll be able to help people understand”.
Ad Revenue Reality Check
Reid addressed the elephant in the room: monetization.
When asked how ad revenue could remain the same with AI Overviews pushing ads lower on the page, Reid explained: “There’s multiple different dynamics going on. On the one hand, sometimes when AI Overviews show up, the ads are lower on the page. On the other hand, the search traffic is just growing overall, right? So we see with AI Overviews people issue more queries, and so at the highest level the two roughly balance out”.
The Ads in AI Mode Story
From earlier reporting, we know:
- Google is beginning to roll out ads in AI Mode Google Search Chief Liz Reid on ads, AI, and the future of search
