Google embeds Gemini across Android as multimodal agent before Apple's WWDC AI reveal

Google is deploying Gemini as an operating-layer AI agent across Android devices, enabling multi-app task completion and screen-aware assistance ahead of its I/O developer conference next week and Apple's expected AI announcements at WWDC.

Core capabilities

According to Google, Gemini Intelligence on Android will:

• Read and understand on-screen content across apps
• Complete multi-step tasks like pulling guest lists from Gmail, generating menus, and populating Instacart shopping carts
• Operate across phones, watches, cars, glasses, and laptops
• Request user approval before completing transactions

Sameer Samat, who oversees Google's Android ecosystem, told CNBC the company is "transitioning from an operating system to an intelligence system," with Gemini serving as the AI foundation. He emphasized that "the human is always in the loop" for agent-driven actions.

Rollout timeline

The app automation features will deploy in phases:

• Summer 2026: Samsung Galaxy and Google Pixel phones
• Later 2026: Expansion to watches, cars, glasses, and laptops

Google is also redesigning Android Auto around Gemini, bringing the assistant to more than 250 million vehicles with what the company claims is its biggest maps update in a decade.

Competitive context

The announcement comes four months after Apple signed a deal with Google to power portions of Apple Intelligence with Gemini. Apple is expected to demonstrate an upgraded version of Apple Intelligence at WWDC, putting pressure on both companies to prove their AI integration strategies.

Google's approach focuses on cross-app coordination and contextual understanding, while Apple has historically emphasized privacy, hardware integration, and user experience control.

Market response

Alphabet's stock has risen more than 140% over the past year as investors respond to its AI strategy, compared to Apple's roughly 40% gain over the same period.

What this means

Google is attempting to establish Gemini as the default AI layer for Android before Apple can demonstrate its own vision for on-device intelligence at WWDC. The "human in the loop" requirement for transactions suggests Google is taking a measured approach to agentic AI, requiring explicit approval for actions that involve spending money or making commitments. The rollout timeline—starting with flagship Samsung and Pixel devices—indicates Google is prioritizing demonstration over immediate broad availability. Whether users will trust an AI to build shopping carts and book reservations remains untested at scale.