Microsoft has recently unveiled a significant upgrade to its AI assistant, Copilot, introducing new capabilities that enable it to perform various online tasks on behalf of users. Activated by "simple chat prompts," these functions operate automatically in the background, allowing users to focus on other activities simultaneously.
The initial batch of supported services includes travel, dining, and gift purchasing sectors. Partner platforms feature well-known names such as Booking.com, Expedia, Kayak, OpenTable for restaurant reservations, and 1-800-Flowers for flower delivery. With Copilot, users can complete restaurant bookings, purchase event tickets, or send gifts to friends. Microsoft notes that this functionality is "compatible with most major websites" and will soon extend to ride-hailing and shared mobility services.
In addition to task execution, Copilot introduces several personalized tools. These include transforming web content into AI-generated podcasts, searching for product discount information, and utilizing the camera to identify objects and respond accordingly. While Microsoft has not disclosed technical implementation specifics, the feature aligns conceptually with similar AI products in the market.
In terms of industry competition, Amazon's Nova Act AI agent, launched in March, already supports browser automation for tasks like web searches and online shopping. Other competitors, including OpenAI’s Operator, Google’s Project Mariner, and Anthropic’s Claude 3.5 Sonnet "computer usage" feature, are still under development or testing. These alternatives primarily emphasize automated clicks and task executions but have yet to be fully rolled out to the public.
Through this update, Microsoft positions Copilot as one of the first AI services offering automated task execution available to the general public. The initial version deployment has commenced, with availability expanding gradually across platforms, markets, and languages over the coming weeks to months. This enhancement could serve as an early benchmark for evaluating the practicality and reliability of AI-powered assistance tools.
