Google's Demand Map: What 14.65 Million AI Conversations Reveal About Where Brands Need to Be

Originally published on The Searchless Journal

Google just published the largest first-party dataset ever on what people actually ask AI. The AI & Economy ATLAS report analyzed 14.65 million conversations from Gemini app, AI Mode, and Gemini API — and found a structural mismatch between how people spend their time and what they ask AI. Government services, professional services (doctors, lawyers, banks), education, and consumer purchases dominate AI conversations at 3x to 20x their share of daily life. This is the friction map, and it reveals where AI creates the most value: in high-stakes, high-friction decisions where people need expert guidance.

But Google shared everything except the data brands need most — whether any of these conversations send traffic to websites. The demand map is visible. The supply side remains opaque. For brands, the strategic implication is clear: if your category sits at the top of the friction map, AI visibility is not optional — it's where your customers are already looking.

The Friction Map: Where AI Over-Indexes

Google's methodology is straightforward. They compared AI conversation volume against the American Time Use Survey (BLS) — how Americans actually spend their days. The ratio of AI conversation share to time share reveals where people use AI to solve problems they can't solve elsewhere.

The findings are stark:

• Government services and civic obligations: AI conversation share is ~20:1 compared to daily time share. People spend almost no time dealing with government in daily life, but they ask AI about government services 20 times more than that baseline suggests.
• Professional & personal care services (doctors, lawyers, banks): >7:1 ratio. High-stakes decisions about health, legal matters, and money dominate AI queries.
• Education: ~6:1 ratio. Learning and skill-building are top AI use cases.
• Consumer purchases: ~3:1 ratio. Shopping and buying decisions are AI-driven at a significant scale.

The pattern is unmistakable. AI is not replacing low-value, routine interactions. It is accelerating high-value, high-friction decisions. The categories where people need the most guidance — health, finance, legal, government — are the categories where AI conversation volume explodes relative to daily life.

The Timing Problem: High-Friction Questions Outside Working Hours

Google's data shows that half of high-friction questions (medical, legal, financial, government) come outside working hours. This is not accidental. People use AI when the alternative is waiting for business hours, booking appointments, or navigating bureaucratic systems that close at 5 PM.

The implications are structural:

• Demand is not seasonal. AI queries for health, legal, and government services happen at 2 AM on Tuesday the same way they happen at 2 PM on Friday.
• The supply side is mismatched. Most human-staffed services close at predictable hours. AI does not. The supply-demand gap is where AI value comes from — and where brands lose if they're not present in those conversations.
• The geography problem. High-friction questions come from time zones where local service providers are closed. AI is global; human support is local.

For brands, this means AI visibility is not a nice-to-have add-on to traditional marketing. It is the only channel that reaches customers when they're actually making decisions — which is often outside the hours your team is working.

What Google Didn't Share: The Traffic Gap

The ATLAS report is comprehensive in every dimension except the one brands care about most: traffic. Google measured conversation volume, topic distribution, and time-of-day patterns. They did not measure how many of those conversations send traffic to websites.

This is not an oversight. It is the defining characteristic of AI search. The question is not "how many clicks does this conversation generate?" The question is "which brands are mentioned, recommended, or trusted in this conversation?" Traffic is a lagging indicator of success in AI search. The leading indicator is mention.

Google knows this. They shared the demand map because they know brands are still operating on a traffic paradigm that doesn't apply to AI discovery. The brands that win will be the ones that stop asking "how do I get clicks from AI?" and start asking "how do I become the brand AI recommends when my customers ask these questions?"

The Category Action Plan: Where GEO Is Existential

The friction map provides a clear framework for AI visibility investment by category. If you're in a high-ratio category, GEO is existential. If you're in a low-ratio category, GEO is optional but still valuable.

High-Ratio Categories (GEO is Existential)

Government & Civic Services (20:1 ratio)

• People ask AI about government processes, regulations, voting, and civic engagement.
• Most government websites are SEO-optimized, not AI-optimized.
• Structured data, API access, and direct knowledge ingestion are the path to visibility.

Health & Medical (7:1 ratio)

• Medical questions, symptom checking, provider recommendations, and health decisions dominate.
• HIPAA and privacy concerns make AI optimization harder — but not impossible.
• Trust signals, expert attribution, and evidence density are the key factors.

Legal & Financial (7:1 ratio)

• Legal questions, financial planning, tax guidance, and regulatory compliance.
• Authority and expertise matter more than freshness.
• Professional certifications, case studies, and regulatory citations are strong signals.

Education (6:1 ratio)

• Learning, skill-building, career guidance, and educational decisions.
• Structured curricula, certification pathways, and expert instructors drive mentions.
• Educational institutions that optimize for AI will dominate student acquisition.

Medium-Ratio Categories (GEO is Strategic)

Consumer Purchases (3:1 ratio)

• Shopping, product research, and buying decisions.
• Product data, reviews, and comparisons are key.
• Ecommerce brands that provide structured product data win.

Low-Ratio Categories (GEO is Nice-to-Have)

Entertainment & Leisure (1:18 ratio or lower)

• Eating, drinking, watching TV, and entertainment activities.
• People spend time here, but AI conversation share is low.
• Optimization is valuable but not existential.

The Strategic Takeaway: Meet Your Customers Where They're Looking

The demand map is the clearest evidence yet that AI visibility is not about chasing clicks. It is about meeting your customers where they're looking for answers — which is in AI conversations, not on your website.

The brands that win will be the ones that:

1. Audit their AI presence in the categories where their customers are asking questions.
2. Optimize for mention, not clicks. Build content, data, and expertise that AI can cite and recommend.
3. Address the timing problem. Be present in AI conversations 24/7, not just during business hours.
4. Invest in the high-ratio categories first. Prioritize GEO investment where the friction map shows the highest demand.

The demand map is now visible. The question for every brand is: are you on it?

Run your AI visibility audit to find out.

Sources

• Google AI & Economy ATLAS v1.0 Report, ai.google/static/documents/GoogleATLASv1.pdf
• Google Research Blog, "Understanding the AI Economy," blog.google/innovation-and-ai/technology/research/understanding-the-ai-economy/
• American Time Use Survey methodology documentation, BLS.gov/tus
• Search Engine Journal, "Google Data Compares Gemini & AI Mode Use Against Daily Life," Matt G. Southern, July 25, 2026

FAQ

What is the AI & Economy ATLAS report?
Google's first comprehensive study of how people use AI, analyzing 14.65 million conversations across Gemini app, AI Mode, and Gemini API to create a demand map for AI discovery.

What does the friction map show?
The friction map shows where AI conversation volume exceeds daily time share. High-ratio categories (government 20:1, health/finance 7:1, education 6:1) are where AI creates the most value because people use AI to solve high-friction problems.

Why didn't Google share traffic data?
AI search doesn't work like traditional search. The question isn't how many clicks a conversation generates, but which brands are mentioned or recommended. Traffic is a lagging indicator; mention is the leading indicator.

Which brands need GEO most?
Brands in high-ratio categories (government, health, legal, finance, education) where AI conversation volume is 3x to 20x daily time share. For these brands, AI visibility is existential.

How do I optimize for AI discovery?
Start with an AI visibility audit to see where you're already mentioned, where you're missing, and what content gaps exist. Then build structured data, API access, and expertise signals that AI can cite.

For more data on AI search trends, see our AI search statistics and our guide to how ChatGPT chooses sources.