AT&T Adds Gemini for Home to Connected Life Security

AT&T has integrated Gemini for Home, Google's generative AI layer for its smart home products, into its Connected Life home security system, according to a company announcement dated November 4, 2026. AT&T states that Connected Life is the first home security product to integrate with Google Home Premium.

The integration brings several Gemini-powered features to the platform. Subscribers can receive Gemini-generated summaries of home events, which are written recaps of what cameras and sensors recorded. They can also run natural-language searches across recorded video history, meaning they ask a plain-English question instead of filtering clips by time or motion zone. Daily summaries of activity inside and outside the home are also available. The Verge reports that Gemini for Home can describe camera-captured events in detail, including who delivered a package and how many items were dropped off.

Voice interaction is part of the package. Connected Life supports Gemini Live on speakers and displays for conversational exchanges, and it provides camera and doorbell insights through Gemini for Google Home. AT&T says the Gemini for Home-powered system cuts through notification overload and makes home activity more relevant and easier to understand.

The Gemini-powered features require AT&T's $29.99 per month professional Connected Life monitoring plan. Customers who want the generative layer therefore pay a recurring monitoring fee on top of the hardware.

AT&T is also adding the ability to arm or disarm the Connected Life system automatically based on proximity to home. The proximity feature is described as something AT&T is adding, not a capability already in service.

AT&T launched Connected Life in December in partnership with Google and Abode. Setup begins with downloading the AT&T Connected Life app from the Apple App Store or Google Play.

The broader context here is a shift in how home monitoring is consumed. A conventional security feed asks the user to watch clips or scroll a timeline. Natural-language search asks the system to index what happened and answer a question about it, which moves the product from alert delivery toward retrieval. For engineers and architects, the open questions sit less in the feature list than in the data path: where camera streams are processed, how long footage and derived summaries are retained, who can query them, and how the system handles misclassified events.

In my view, the monitoring-tier requirement deserves attention alongside the feature set. Tying generative capability to a recurring plan makes the subscription the route to the feature, which is a different proposition from a one-time hardware purchase. Buyers evaluating the offer will want to weigh the monthly cost against the capabilities they will actually use.

The practical upside is easy to picture. A household fielding dozens of motion notifications a day gains a way to ask whether a delivery arrived or when the back door last opened, and to get a direct answer. If the proximity-based arming arrives as described, the routine of setting and clearing the system could become less of a chore. Whether that convenience holds up depends on the retention and access controls that accompany the summaries, and those details will matter more than the headline features as the product matures.