Microsoft working on voice avatars, generative layouts, and agents for Copilot

Microsoft is actively developing several new features for Copilot, including interactive AI characters, UI enhancements, and generative capabilities. These updates aim to make Copilot more dynamic, visually rich, and conversationally powerful.

Revamped UI: New Prompt Bar and Enhanced Navigation

One of the biggest UI changes in development is a redesigned prompt bar, also known as the composer UI. The updated version will resemble a widget with improved shadowing and visual cues to indicate when users are conversing with a specific AI character. For instance, it may display a message like “Talk to Aqua” when engaging with a character-based AI assistant.

Another major improvement is a new toggle that lets users switch between “Quick Responses” and “Think Deeper” modes, offering clearer distinctions between fast, concise answers and more in-depth AI-generated insights. Compared to the current UI, this feature is expected to be more intuitive and permanent.

Meanwhile, the side navigation bar has started displaying chat history properly, suggesting that this feature is nearing an official release. Additionally, a Discover button has appeared at the top, potentially leading to a more streamlined way of exploring Copilot’s capabilities.

Daily Briefing and Conversational Enhancements

Microsoft is also improving the Daily Briefing feature, which provides AI-generated summaries of relevant information. A new capability under development will allow users to open a chat directly from the briefing and ask Copilot follow-up questions about the topics mentioned.

Although this feature is not yet live, it signals Microsoft’s push to make AI more interactive and context-aware in everyday workflows.

New Experimental Features: Avatar and Generative Layout

Microsoft Labs has introduced two new experimental features that could significantly impact how users interact with Copilot.

Avatar – Voice Mode Upgrade

The first experiment, Avatar, allows users to “Talk with an avatar in voice mode.” While details remain unclear, it could mean the introduction of AI-powered avatars in Copilot’s voice UI. Alternatively, it might expand upon the previously discovered AI characters.

At present, enabling this feature breaks the existing voice UI, suggesting that it is still in an early development phase but intended specifically for voice interactions.

Generative Layout – AI-Powered Visuals

The second experiment, Generative Layout, offers an exciting new way for Copilot to provide answers. It is described as a feature that allows users to “Go beyond text with generated visual answers.”

This strongly suggests that Copilot will soon be able to generate images alongside its text-based responses, similar to OpenAI’s upcoming advancements in multimodal image output.

Recent leaks indicate that ChatGPT is also preparing to support multimodal image generation, a long-awaited feature from OpenAI. While OpenAI has yet to release this update, Sam Altman recently hinted that it is coming relatively soon. Given the history of OpenAI and Microsoft’s partnership, these image-generation capabilities could launch simultaneously on both ChatGPT and Copilot, or even appear in Copilot first.

Future of Copilot: AI Agents on the Horizon?

One of the most intriguing discoveries in Copilot’s development is the mention of agents in Microsoft’s internal code. While details are scarce, this suggests that Copilot could soon introduce agentic AI functionalities, allowing users to automate tasks or interact with more autonomous AI-driven systems.

When Will These Features Be Available?

As of now, none of these updates are publicly available, and Microsoft has not provided an official timeline for their release. However, given the increasing visibility of these features in Copilot’s internal testing, it’s likely that some of them will roll out in the near future.