Microsoft is enhancing its AI assistant, Copilot, with new “vision” capabilities. This feature, initially announced in October, is now being previewed with select Copilot Pro subscribers. These users can activate Copilot Vision within their Edge browser to interact with and analyze the content of webpages.
How Copilot Vision Works
Copilot Vision acts as a browsing assistant, residing at the bottom of the Edge browser. Users can engage Copilot Vision to analyze webpages, including text and images, and assist in decision-making. This feature aims to streamline tasks like research and shopping by providing a “second set of eyes” to quickly scan and analyze information.
Privacy and Security Measures
Microsoft assures users that all context and information shared during a Copilot Vision session is deleted upon session closure. Additionally, website data is not retained for model training purposes. The Copilot team emphasizes its commitment to copyright protection, creator rights, user privacy, and safety.
Current Availability and Future Plans
Currently, Copilot Vision is available to a limited number of Copilot Pro subscribers in the US who have opted into the early-access Copilot Labs program. The feature is opt-in and initially functions with a select group of websites. Microsoft plans to gradually expand the capability based on user feedback, eventually supporting more Pro users and websites.
Potential Expansion and Competitive Landscape
While there is no official confirmation, Microsoft may extend Copilot Vision to other products like OneDrive and Excel. This expansion could significantly benefit enterprise users by simplifying workflows and decision-making processes. Microsoft’s cautious approach suggests that wider integration may take time.
The launch of Copilot Vision’s preview coincides with advancements in agentic AI by competitors like Salesforce and Anthropic. Salesforce has introduced AgentForce to automate workflows, while Anthropic’s “Computer Use” allows developers to integrate their AI model, Claude, for more complex interactions with computer desktop environments.
Add Comment