Mustafa Suleyman, co-founder of DeepMind (now Google DeepMind) and Inflection AI, was appointed as the CEO of Microsoft’s AI division in March 2024. Since then, the Copilot experience has seen a noticeable decline. However, after six months under Suleyman’s leadership, Microsoft has finally refreshed Copilot with an updated look.
This marks a notable departure from the previous user interface, which adhered to Microsoft’s traditional design language. The new Copilot showcases a card-based user interface with rounded corners and high-quality visuals on its homepage. During my brief usage, Copilot felt more personal and approachable, creating an inviting and friendly atmosphere.
If you’re wondering where you’ve seen this interface before, Suleyman’s previous startup, Inflection AI, has a remarkably similar design in its Pi chatbot (visit). The Pi chatbot emphasizes personal AI, a quality that is also evident in the revamped Copilot.
You can now engage in voice chats with Copilot Voice, which promises a natural interaction similar to ChatGPT’s Advanced Voice Mode. I attempted to communicate with Copilot Voice, but it was currently experiencing heavy traffic.
Moreover, Microsoft introduced Copilot Vision, which enables users to browse the web while interacting with Copilot. It can see what’s on your screen, allowing you to chat with it and request suggestions, ideas, and more. Microsoft states that this is an experimental feature and will only be available to Copilot Pro users under Copilot Labs.
Copilot Vision is set to be integrated into the Edge browser, in the starting will function on select websites. It will also be an opt-in feature, allowing users to choose whether they want to share their screen with Copilot Vision. Following the Windows Recall fiasco, Microsoft assures users that Copilot Vision will process all data locally on the device.
In addition, Microsoft has introduced a feature called “Copilot Daily,” which serves as an AI-powered news presenter. Users can listen to global news directly on Copilot’s homepage, featuring an AI-generated voice. There are four available AI voices: Canyon, Grove, Meadow, and Wave.
As for the OpenAI model powering the new Microsoft Copilot, it’s hard to determine at this point. However, during my brief testing, it accurately answered all the complex reasoning questions I posed, which is quite impressive. Unlike OpenAI o1 models, it doesn’t take much time to “think”; instead, it provides the correct answer on the first try. I’m genuinely impressed by the AI model behind the new Copilot.
Microsoft also mentioned that Copilot will soon introduce an experimental feature called “Think Deeper” as part of the Copilot Labs offerings. This feature will take some time to process complex queries, similar to OpenAI’s earlier models. Same as of Copilot Vision, this feature will be available exclusively to paid Copilot Pro users.
Have you had a chance to try the revamped Copilot yet? Visit copilot.microsoft.com or download the Copilot app on Android or iOS, and share your thoughts with us!
0 Comments