Google’s Gemini Spark: a 24/7 AI agent that plans complex tasks autonomously

At I/O 2026, Google introduced Gemini Spark — a 24/7 personal AI agent designed to autonomously plan and execute complex, multi-step tasks across applications. Unlike traditional chatbots that react to single prompts, Spark leverages the new Gemini 3.5 Flash model and Google's Antigravity framework to maintain persistent context, decompose high-level goals into subtasks, and carry them out across Google Workspace, Chrome, and third-party APIs without user supervision. The agent can, for example, research a topic, draft a report, schedule a meeting, and send follow‑up emails — all in one continuous workflow.

Starting next week, Gemini Spark will be available exclusively to Google AI Ultra subscribers in the United States, priced at $100 per month. The initial rollout includes deep integration with Google’s ecosystem (Gmail, Calendar, Docs, Drive, Maps), with Chrome extension and third‑party app support promised in later updates. By making an always‑on, task‑orchestrating agent accessible to professionals, Google is positioning Spark as a direct competitor to OpenAI’s upcoming Operator and Anthropic’s computer‑use agents, aiming to redefine how users delegate daily productivity work.