Google is rolling out information agents in AI Mode, introducing a feature that runs in the background to track topics across the web and send users synthesized updates when something changes.
According to Google’s May 19 product blog and contemporaneous reporting cited in the source material, the feature is the first publicly available part of the company’s broader Search agents effort introduced at I/O 2026. It is available now to AI Ultra subscribers, with multiple outlets reporting that AI Pro subscribers are expected to get access later this summer.
Background monitoring inside AI Mode
Reporting cited in the source says information agents continuously watch a range of sources, including blogs, news sites, social posts, shopping, finance, and sports feeds. When the system finds relevant updates, it sends push notifications and provides a synthesized summary.
The source material describes the setup as an agentic extension of Search that combines ongoing monitoring with multi-source synthesis. It also says the rollout is tied to other Search changes Google outlined in its product blog, including an "intelligent Search box" and a new Flash model aimed at agentic tasks.
sbb-itb-dc5b06c
How the feature works
As described in reporting from TechCrunch and Android Authority cited in the source, users can create an active tracked topic in AI Mode by entering a natural-language request such as "keep me updated on X". That tracked topic then appears in AI Mode history, and the Google app sends notifications when the agent identifies relevant developments.
The source also says the redesigned AI-powered Search box supports longer, multimodal prompts.
Availability and pricing
Multiple outlets cited in the source report that information agents are already live for AI Ultra subscribers. The same reporting says the feature is set to expand to AI Pro subscribers this summer.
The source notes that 9to5Google lists AI Ultra at $99.99 or $199.99 per month.
What observers are watching
The source says several open questions remain around the rollout, including which subscription tiers and markets receive the feature and when, how often notifications are sent, and what source coverage and provenance signals users will see.
It also identifies practical measures to watch, including false-positive notification rates, latency for time-sensitive information such as finance and ticket drops, and the controls users have to manage and review active agents.
Broader significance in the source material
The source frames persistent, goal-directed agents as a growing trend in consumer AI and says coverage has positioned Google’s information agents as a step beyond scheduled alerts toward continuous, model-backed monitoring and synthesis.
For practitioners, the source highlights sustained background monitoring, multi-source synthesis, and the integration of the feature into a product that, according to Google’s blog as cited in the source, reached more than one billion monthly users within a year.
The source also says the rollout offers a public example of how a major platform may deliver continuous monitoring and synthesis as a consumer-facing feature, while raising familiar operational issues around rate limits, content provenance, and real-time connector reliability.
The source adds that the available reporting covers feature availability, interface behavior, and example use cases, but says additional technical documentation or user-control details would be needed to assess reliability and privacy implications in greater detail.
