Microsoft reportedly working on 3D model maker and voice avatars for Copilot

Microsoft Copilot is discreetly working to expand the capabilities of its AI platform. Recent developments hint at the fact that the tech giant is working on animated voice avatars and image‑to‑3D object generation.  Both upgrades were spotted in unreleased builds by TestingCatalog. While they are clearly labeled “Labs,” they provide an early look at features that could make Copilot feel a lot more personal and a lot more useful.

The first update is a change to the Characters feature tied to Copilot’s voice mode. The toggle now has a “Labs” tag, indicating that this is still an experimental feature. The character avatars are meant to make voice interactions with Copilot more personalised, but the update only involves small UI tweaks. There is however, no official release timeline, and the feature remains limited to test builds for now.

Microsoft is also experimenting with Copilot 3D, which, as the name suggests, is a tool designed to generate 3D models from user-uploaded images. The interface now includes categories like characters, scenes, and objects, along with options to rotate, zoom, and preview models in grayscale or color. Since this feature is also in explorer mode, it allows users to remix existing 3D figures using prompts.

Another experimental feature is to add playback support for Copilot Actions. It allows users to replay browser automation sessions. This could be useful for reviewing or adjusting automated tasks, especially for testers or power users.