Google unveils 3 new AI tools to automate commerce

Google is betting that the future of artificial intelligence lies not in answering questions, but in completing tasks. At its I/O 2026 conference, the company unveiled a suite of new products centered on Gemini Spark, a persistent AI agent designed to work across Google’s services even when the user is offline, marking its most ambitious push yet into the world of autonomous systems.

"We are in that part of the cycle where people want to see real value in the products they use on a day-to-day basis," Sundar Pichai, CEO of Google and Alphabet, said in a briefing. He argued that with Spark, value comes from an agent that operates around the clock in Google's cloud, so "you don't need to keep your laptop open to make sure it's running."

The agent is powered by Google’s new Gemini 3.5 Flash model and runs within a framework called Antigravity, allowing it to execute multi-step workflows across applications like Gmail, Docs, and Calendar. The service will be available to subscribers of a new Google AI Ultra plan, which starts at $100 per month for five times the usage limits of the Pro plan and 20TB of storage.

The announcements place Google in a fierce race against Microsoft, OpenAI, and Anthropic to build the definitive AI agent. The company is wagering that its deep integration with consumer services and a massive $180 billion to $190 billion capital expenditure plan for the year will give it a decisive edge in a market shifting from conversational AI to autonomous action.

Protocols for an Agent-Driven Economy

For agents to become truly useful, they need to interact with the real world, and particularly with commerce. Google addressed this by introducing two new protocols: the Agent Payments Protocol (AP2) and the Universal Commerce Protocol (UCP). AP2 is a system designed to let AI agents make secure purchases within boundaries defined by the user, with Google comparing the safeguards to giving a teenager their first debit card.

The system creates a verifiable link between the user, merchant, and payment processor. It is built on UCP, an open-source standard announced earlier this year that gives agents and commerce systems a common language. The UCP Tech Council now includes rivals like Amazon, Meta, and Microsoft, suggesting a broad industry consensus on the need for a shared framework for agent-driven commerce. Google is also rolling out a Universal Cart this summer that works across Search, Gemini, YouTube, and Gmail to track prices and deals.

New Tools Target Creators and Consumers

Alongside its agent-focused announcements, Google also launched tools aimed at making AI more accessible for content creation. Google Pics, a new design and image-generation app, is positioned to compete directly with Canva and Anthropic's Claude Design. Powered by the Nano Banana 2 model, Pics allows users to generate and edit complex graphics using simple text prompts or by leaving comments directly on image elements, similar to Google Docs.

For video, Google is integrating its Gemini Omni model into YouTube Shorts. The feature allows creators to generate and remix videos using text, images, and audio prompts. The system is designed to handle complex edits automatically while preserving character consistency across scenes. YouTube emphasized creator protections, including watermarks and opt-out controls for AI-generated remixes.

The Race for the Definitive AI Agent

Google's strategy with Spark is distinct from its competitors. While OpenAI's agent primarily uses a browser and Anthropic's Claude Cowork operates on a user's desktop, Spark works through structured API integrations with Google's own Workspace apps and third-party services. This approach promises greater reliability and speed than screen-reading, but it also means Spark's capabilities are initially confined to the ecosystems Google is connected to.

The reliability of these autonomous systems remains a critical challenge for the entire industry. An agent that misinterprets a user's intent could have significant consequences, a risk Google is mitigating at launch by requiring explicit user approval for any financial transaction. The real test for Gemini Spark isn't whether it can do the work, but whether users will trust it enough to let it.

This article is for informational purposes only and does not constitute investment advice.