Three months after OpenClaw generated buzz as an "always-on" AI agent, Google launched its own answer: Gemini Spark, announced at Google I/O 2026.
Spark runs in the background rather than waiting for you to open an app and ask it something. According to Google's announcement, it can write emails on your behalf, create study guides that update automatically as new information becomes available, and scan credit card statements for subscription fees you may have forgotten about.
This positions Spark closer to an autonomous assistant than a chatbot. The distinction matters: most AI tools today are reactive - you type, they respond. An always-on agent acts without being prompted, which is more useful in theory but also raises real questions about what it's accessing and when.
The subscription monitoring is the most concrete use case Google described. Forgotten or hidden subscriptions are a genuine problem - research consistently shows people underestimate their monthly subscription spend by $100 or more - so an agent that flags them automatically has clear practical value.
The email drafting capability is less novel; Gmail already offers AI-assisted drafting. The "continually updated study guides" feature sounds promising but the announcement was light on specifics. It's unclear whether Spark is synthesizing new information as it appears or simply appending to an existing document.
Gemini Spark is Google's direct answer to the agentic direction Anthropic has been taking with Claude and what standalone agent platforms have been pitching for the past year. The race to build AI that works in the background - rather than tools you have to actively pull up and prompt - is becoming the central competition in the assistant space. Whether Google can execute it at a quality level that makes people actually trust an agent to act on their behalf is a separate question from whether the product exists.