Google on Monday unveiled Gemini Spark, a general-purpose AI agent that runs 24/7 on Google Cloud infrastructure — drafting emails, assembling documents, monitoring inboxes, and eventually making purchases — even after you close your laptop and lock your phone. Announced at Google I/O 2026, Spark is Google's most aggressive move yet to transform its AI assistant from a tool that answers questions into one that autonomously completes work.
What Gemini Spark actually does
The architectural distinction matters. Unlike conventional AI assistants that activate when prompted, Spark runs persistently on Google Cloud, powered by the new Gemini 3.5 Flash model and Google's Antigravity agent orchestration platform. You give Spark a complex instruction — "email my boss a status update pulling the latest figures from our shared spreadsheet and the project timeline in our Slides deck" — and it executes across multiple Google applications without further input.
"When you use it, it almost feels like you're tossing things over your shoulder — Spark's catching them and gets the job done," said Josh Woodward, VP of Google Labs, Gemini App, and AI Studio, during a press briefing ahead of the keynote.
In practical terms, Spark can set triggers to monitor credit card statements for hidden subscription fees, watch your inbox for updates on a specific project and compile a daily digest, or assemble raw meeting notes from emails and chats into a polished Google Doc. According to SiliconANGLE's reporting, these aren't hypothetical — they reflect how early testers have actually been using the product.
Third-party integration is coming fast. Model Context Protocol connections to Canva, OpenTable, and Instacart launched immediately, with Google planning more than 30 third-party partners in the coming months. Future capabilities include texting and emailing Spark directly, creating custom sub-agents for specialised tasks, and operating the local browser.
The teenager with a debit card
For all its ambition, Spark faces the same fundamental challenge as every AI agent: trust. Particularly when money is involved.
When asked during the press briefing how Spark would avoid making unauthorised purchases, Woodward was strikingly candid. "On the team, we think a lot of it is like if you're giving a teenager their first debit card — there's sort of limits and sort of constraints around it, and that's how we'll be designing Spark as we go through the year," he told VentureBeat.
At launch, Spark won't autonomously make purchases. But Google has built the infrastructure for that future. The company introduced the Agent Payments Protocol (AP2) — a system that lets AI agents make secure purchases within user-defined boundaries. Users set specific brands, products, and spending limits. If criteria are met, the agent can complete a purchase. AP2 generates a permanent digital paper trail using tamper-proof digital mandates.
This connects to a broader trend we've been tracking. When Cloudflare and Stripe launched an open protocol for AI agent commerce earlier this month, it signalled that the industry is building payment rails specifically for autonomous agents. Google's AP2 — backed by a Universal Commerce Protocol council that now includes Amazon, Meta, Microsoft, Salesforce, and Stripe — is the most comprehensive version of this idea yet.
What this means if you run a business on Google Workspace
If your organisation runs on Google Workspace, Spark is the most significant Gemini announcement in the past year. It's the difference between an AI that helps when you ask and one that works while you don't.
The consumer product starts at $100/month on the Google AI Ultra tier, which includes a 5x higher usage limit, priority access to Antigravity, and 20TB of cloud storage. For enterprise customers, Google's Gemini Enterprise Agent Platform starts at $30/user/month (with a Business tier at $21/user/month), offering no-code agent builders, prebuilt agents for software development and customer engagement, and connections to both Google and Microsoft 365 apps.
That $30/user figure puts Google in direct, price-matched competition with Microsoft Copilot, which recently passed 20 million paid enterprise seats at the same price point. The competitive reality is simple: most businesses will default to whichever agent integrates with their existing productivity suite. If you're a Google Workspace shop, Spark is your path. If you're on Microsoft 365, Copilot is.
But the always-on architecture is a genuine differentiator. Microsoft Copilot has traditionally been prompt-activated — you ask, it does. Spark is designed to run continuously, monitoring and acting without being asked. Microsoft has responded with Copilot Cowork, which also runs in the cloud, but Spark's deep integration with Gmail, Docs, and the broader Google ecosystem (Search, YouTube, Chrome) gives it a broader surface area than any competitor.
For Australian businesses specifically, this builds on a pattern we've tracked closely. Macquarie Bank recovered 130,000 hours in seven months by deploying Gemini Enterprise to all 5,000 employees, and Google's Chrome AI Skills feature already brought no-code Gemini automation to the browser. Spark is the next logical step — but a much bigger one.
The five-way race to build the agent that does your job
Google is not building in a vacuum. The announcement lands in the most intense competitive period in AI agent history, with five major platforms placing fundamentally different bets:
OpenAI recently unified its Operator and deep research capabilities into ChatGPT agent, which operates primarily through a virtual browser. Anthropic launched Claude Computer Use and Claude Cowork, which work directly on your desktop by seeing and controlling your screen. Microsoft introduced Copilot Cowork tightly bound to the Office 365 ecosystem. Apple is preparing a revamped Siri for WWDC 2026 as an always-on agent — reportedly powered by Google's own Gemini models through a deal costing Apple around $1 billion per year.
Google's bet with Spark is distinctive: rather than controlling your screen pixel by pixel like Anthropic, or operating through a browser like OpenAI, Spark works through structured integrations — Google's own Workspace APIs and third-party MCP connections. The advantage is reliability and speed. The disadvantage is that Spark can only act within the systems it's been connected to.
What to watch
Spark begins rolling out to trusted testers this week, with a broader beta for U.S. Google AI Ultra subscribers next week. Three things will determine whether this matters beyond the announcement:
Reliability at scale. Google's Gemini app has 900 million monthly users, but the history of digital assistants — from Clippy to early Siri to Alexa — is littered with products that promised proactive intelligence and delivered frustration. An agent that emails the wrong person or misreads a spreadsheet figure creates consequences that are difficult to reverse.
Third-party ecosystem. Spark's utility depends on how quickly MCP integrations expand beyond the initial 30 partners. The more business tools Spark can connect to, the more useful it becomes — and the harder it becomes for businesses to leave Google's ecosystem.
Australian availability. The beta is U.S.-only. For Australian businesses watching this space, the timeline for local availability — and whether enterprise features will be priced differently for the Australian market — remains unclear.
Sundar Pichai framed the opportunity in economic terms during the keynote. Companies processing roughly one trillion tokens per day on Google Cloud could save over $1 billion annually by shifting 80% of workloads to a mix of Flash and frontier models. "You no longer have to trade quality for latency," Google said. Whether businesses will trade control for convenience — that's the question Spark will have to answer.
Sources
- Google unveils AI model Gemini 3.5 and AI agent Gemini Spark — CNBC
- Google's new Gemini Spark is an always-on AI agent for daily digital tasks — SiliconANGLE
- Google's new AI agent can draft your emails, monitor your inbox and eventually spend your money — VentureBeat
- Google I/O 2026: Defining the agentic AI era — Google I/O
- Google announces slew of AI advances, including a personal AI assistant — AP News
- Gemini Enterprise Agent Platform pricing — Google Cloud
