Our Take
Google is shipping agentic Search as a subscription feature, not as a free upgrade—the real product is capturing user intent before it leaves the Search box.
Why it matters
Search agents represent the first mainstream deployment of autonomous monitoring at consumer scale. This shifts Google's monetization model from answering queries to selling ongoing task automation, and it pressures competitors to match agentic capabilities in their own search products.
Do this week
Product leads: audit your search or browsing UX to identify tasks users repeat weekly—those are the use cases Google's agents will cannibalize this summer.
Google Search Now Deploys AI Agents That Monitor Tasks for You
Google announced today that it is integrating AI agents directly into Search, starting with information agents that run in the background to monitor for changes on the web. Users can define criteria (apartment location, price range, amenities; sneaker release drops; job postings in a specific city) and the agent will scan blogs, news sites, social posts, real-time finance and shopping data, and sports feeds, then send synthesized updates when matches appear.
Information agents will launch first for Google AI Pro and Ultra subscribers this summer, then expand based on uptake. Google is also expanding agentic booking capabilities to local services (karaoke rooms, home repair, beauty, pet care), with the ability to initiate phone calls on behalf of users to complete transactions—rolling to all U.S. users this summer.
The announcement includes three additional capability expansions: generative UI in Search (custom dashboards, trackers, and mini-apps built on the fly using Gemini 3.5 Flash); expanded Personal Intelligence (connecting Gmail, Google Photos, and soon Google Calendar) to 98 languages and 200 territories, free; and an overhauled Search box described as its largest redesign in 25 years, with AI-powered suggestions and multi-modal input (text, image, file, video, Chrome tab).
Gemini 3.5 Flash is now the default model in AI Mode for all Search users globally (company-reported). AI Mode itself has reached over one billion monthly active users, with query volume more than doubling each quarter since launch (company-reported).
Agents Turn Search Into a Subscription Retention Product
This moves Google Search away from transactional queries (answer this question now) toward task automation (monitor this for me indefinitely). That shift unlocks recurring subscription revenue and stickiness: a user paying for an apartment-hunting agent returns daily regardless of whether new listings appear.
The 25-year Search box redesign is cosmetic cover for a deeper structural change. The real product is capturing high-intent user queries before they leave Google—booking a karaoke room, finding a contractor, hunting for a job—and fulfilling them inside Search rather than sending users to third-party sites. Each booking agent call keeps transaction data, user preferences, and search behavior inside Google's walled garden.
For competitors, the timing is urgent. Agents require two things: continuous model execution (not one-shot queries) and real-time data access (web, commerce, calendar, location). Neither Microsoft Copilot nor OpenAI's search product has announced comparable monitoring capabilities at scale.
What Builders Should Do Now
If you ship a product where users repeat the same search weekly (job boards, real estate platforms, price tracking, calendar integration), Google's agents will compress your traffic. Audit which user workflows depend on recurring visits to your site and begin migrating those workflows into your own API or native app before Google's agents launch broadly.
If you integrate with Google Search (via shopping feeds, local business profiles, or web indexing), expect agent queries to spike volume and reduce click-through to your site. Prepare your backend to handle sustained, non-human traffic from Google's 24/7 monitoring.
For LLM application builders: agents that monitor external state asynchronously are now a reference implementation at scale. Google's design (users define intent in natural language; agents reason across multiple data sources in real time; updates arrive via notification) is the pattern the market will adopt.