The Future with Agentic: Google's Shift from Information to Intelligence
Summary
Dr. Stefan Ebener, Head of Customer Engineering at IMIA Google, delivered a keynote at ITB Berlin 2026 titled 'The Future with Agentic: Google's Shift from Information to Intelligence,' moderated by Dirk Rogl, Managing Director of Travel Commerce. The session argued that the internet is undergoing a fundamental transformation—away from click-based information retrieval toward autonomous, goal-driven AI agents that discover, evaluate, and purchase on behalf of consumers. Ebener opened with a striking claim: 'The internet as you all know it today will dramatically change and that happens right now. The times where you click a button to seek something or to purchase a product might be gone in a year from now.'
On the search side, Ebener cited 5 trillion annual search queries at Google and highlighted that queries of five or more words are growing 2–3x compared to simple keyword searches, driven heavily by the 18–35 demographic. The shift from keyword to contextual, conversational queries (e.g., 'best spot on Bali with my three kids in January') represents a fundamentally different mode of consumer intent expression that powers Google's new AI Mode, which combines image search with shopping data for an inspirational and visual discovery experience.
Google Lens was presented as a key agentic signal channel: 20 billion visual searches per month, with 1 in 5 having explicit purchase intent. AI Overviews now reaches 1.5 billion users worldwide and is positioned as a critical visibility layer for travel brands.
A major policy announcement was the Universal Commerce Protocol (UCP), introduced in the US in January 2026, an open-source agentic protocol connecting merchants and consumers for end-to-end hyperpersonalized commerce—including seamless checkout and payment. Ebener acknowledged European rollout timing remains undisclosed but confirmed it is coming. The audience showed high awareness of UCP, surprising Ebener.
A live demo using a real retail web shop illustrated agentic commerce in practice: a user requested a birthday party package for a 7-year-old who loves dinosaurs, for 7 kids and 3 adults the following week. The agent surfaced curated options, handled a budget constraint (reducing total from €472 to €250 by reasoning about necessity vs. non-necessity of items), offered upsell/cross-sell suggestions, and completed a seamless checkout via Visa integration—all conversationally, without manual browsing.
Ebener drew a direct parallel to the travel industry: the same technology that enables virtual clothes try-on (already live in Germany via Binga) can deliver virtual destination try-outs—immersive multi-dimensional views of resorts, neighborhoods, and landmarks under different seasonal conditions, embeddable on travel websites.
On SEO vs. GEO, Ebener declared: 'The times where you only concentrate on SEO is over. You also need to have a focus on GEO—Generative Engine Optimization.' He stressed that context engineering has replaced prompt engineering: high-quality images, precise product descriptions, and clear value propositions are the new ranking signals for LLMs like Gemini. Wayfair was cited as a case study using Gemini to improve product descriptions and images, resulting in significantly higher conversion rates and cost savings.
The Q&A surfaced a key industry tension: travel is acknowledged to be more emotionally complex and high-stakes than retail. Ebener conceded that if a wrong product arrives (e.g., blue shoes instead of red), it is easily returned, but if an AI misclassifies a resort as family-friendly when it is not, the consequences are far more damaging—placing a higher burden on data quality and precision for travel providers.
Key Takeaways
- •Ebener predicted the end of click-based internet within one year: 'The times where you click a button to seek something or to purchase a product might be gone in a year from now,' framing the shift to agentic AI as a fundamental, not incremental, change.
- •Google processes 5 trillion search queries annually; queries of 5+ words are growing 2–3x faster than keyword searches, driven by the 18–35 demographic—signaling that consumer intent expression is becoming richer and more contextual.
- •Google Lens reaches 20 billion visual searches per month, with 1 in 5 having purchase intent—making it a high-signal, underutilized channel for travel brands to capture inspiration-to-booking conversions.
- •AI Overviews has 1.5 billion users worldwide, representing a massive new visibility surface that travel brands must optimize for, separate from traditional search rankings.
- •Google introduced the Universal Commerce Protocol (UCP) in the US in January 2026—an open-source agentic protocol enabling end-to-end commerce from discovery through seamless checkout; European rollout is planned but no date confirmed. Ebener's cue that Euro currency display signals readiness drew knowing laughter.
- •A live demo showed an agentic system planning a dinosaur-themed birthday party (7 kids, 3 adults, next week), handling a budget reduction from €472 to €250 by reasoning about necessity of each item, offering upsells, and completing Visa checkout—all conversationally in real time from a real retail site.
- •Autonomous price-triggered booking was cited as a core agentic use case: 'Whenever a flight to Bali drops below 600, that flight should be automatically booked for me'—even while the consumer is asleep, with no human intervention required.
- •Wayfair case study: used Gemini to improve product descriptions and images at scale, achieving 'massively increased conversion rates' and significant cost savings—directly applicable to hotel and resort catalog optimization in travel.
Transcript
When AI becomes a relevant player in travel, who's taking over the control? Here is the answer. Here's the AI track. Welcome ladies and gentlemen. Good to have you here. I'm Duke Hogal, managing director of travel commerce. And within the next three hours, we talk about surprise. We talk about AI and we try to avoid buzzwords and we try to to think accurate on time. So far AI and travel has been framed as a tool. It was about better search, smarter pricing, more efficient operations. But this ye...
