How Microsoft's new AI Copilot features could transform teamwork and projects

Build 2024: Microsoft's new AI tools aim to enhance team collaboration, automate processes, and improve productivity. Here's how.
Written by David Gewirtz, Senior Contributing Editor
Microsoft Copilot logo outside of Build event
Sabrina Ortiz/ZDNET

Microsoft has been working hard to integrate artificial intelligence into all its offerings and, for the most part, it has succeeded. Most of its integrations are provided under the overarching brand name of Copilot. There are Copilots in Word, Excel, and PowerPoint, in industry solutions like finance and government, in Microsoft's security offerings, and even as part of Microsoft Visual Studio and GitHub development resources.

For as long as the concept of artificial intelligence has existed in the minds of forward-looking computer scientists, the idea of intelligent agents has also existed. The premise is that these agents could accomplish tasks that are otherwise tedious or annoying, and do so intelligently, thus freeing up the time of professionals to provide their unique value-added capabilities to their organizations.

Also: I went hands-on with Microsoft's Surface Pro and Laptop, and I'm ready to ditch my MacBook

Microsoft may be achieving some of those dreams here at Build 2024. At this event, Microsoft is announcing a variety of intelligent agent and AI-based automation tools to increase productivity on multiple levels. Some of what Microsoft is announcing is very preliminary and may not even enter preview for a while. As such, the details of what Microsoft is announcing are somewhat vague.

That said, here's what we know.

Team Copilot

Let's start off with a clarification. Team Copilot and Microsoft Teams are different entities. Team Copilot works with actual teams (lowercase "t") of people. It's used within Microsoft's Teams, Loop, Planner, and Microsoft 365 applications.

Microsoft says Team Copilot "goes beyond a personal assistant," and can help groups at the team, department, or full company level. The intent is to help team management be more productive and collaborative.

Also: Every Copilot+ PC Microsoft just announced to take on Apple's M3 MacBooks

To this end, Team Copilot provides three primary services.

The first is Meeting Facilitator. This tool takes notes in a meeting and manages meeting agendas. It's not clear whether the note taking is just a form of transcription, or if the tool adds value by summarizing and organizing the meeting notes. At this stage, it's too early to know whether "manages meeting agendas" means coordinating meeting times, or acting as a communications intermediary to get agreements on actual meeting topics.

Next up is Group Collaborator, which is designed to help everyone get more out of chats. Since Microsoft Teams has individual channels, it serves as something of a combination of Zoom and Slack. Group Collaborator adds to those capabilities, possibly by monitoring chat channels.

From the way Microsoft describes it, it looks like it will spotlight the most important information, identify action items throughout the discussion, and find and help manage unresolved issues. We've so far only been given one sentence to work from, but it looks like the AI will have context awareness of the flow and structure of chat-based meetings.

Also: Yikes! Microsoft Copilot failed every single one of my coding tests

The last main component of Team Copilot is Project Manager. Microsoft says Team Copilot "ensures every project runs smoothly." It creates and assigns tasks, tracks, deadlines, and notifies team members when input is needed. Of course, Microsoft has long had high-quality project management software, which also does task and deadline management. Reading between the lines, I'm guessing that the Team Copilot version does this work by processing Teams chats and other documents to provide some form of intelligently helpful guidance.

These capabilities will be in preview later this year.


One of the Copilot-related offerings that I find particularly interesting is the company's Copilot Studio. This is a no-code or low-code interface for building custom copilots based on your own business data.

For example, you might upload a set of PDFs that contain your HR policies into Copilot Studio, and then enable generative AI to answer questions based on those documents. This becomes particularly interesting when you can embed a custom copilot, based on these documents, right onto your company's website.

Copilot Studio has a wide range of generative AI capabilities as well as a great many plugins that facilitate bringing data into and out of a given copilot. This would allow you to set up a copilot to track revenue, cross-index customer data, and deliver on a wide range of business information needs.

Also: These four new Copilot for Microsoft 365 features make prompt writing like a pro even easier

Agents take the custom copilot idea to the next level. Now, instead of having those interactions driven solely by an individual, agents can do work dynamically. Microsoft says that agents can automate long-running business processes, reason over actions and user inputs, leverage memory to bring in context and learn based on user feedback.

The company describes one example, of an "order taker" copilot that manages end-to-end order fulfillment processes, from order taking, to processing orders, to managing and substituting out-of-stock items.

One idea that came to me was a copilot that drives the collections process. Accounts receivable operations have long had armies of collections agents making phone calls, and some simple automated processes which send dunning letters. But with an automated copilot, it could be possible to have full conversations with the financed party, listen to issues raised, conduct negotiations, offer compromises, and more. Conversations could be initiated via an outgoing text message, and then conducted entirely over text interaction.

This could reduce the staffing requirements considerably, which could be a win for lenders. Of course, any time we replace jobs with machines, there are some serious societal concerns. Even so, it's up to each company to balance cost reduction with quality of service. An AI agent might make accessing support easier for borrowers, and allow the company using the AI to allocate its more experienced personnel to problems more worthy of their attention.

The ability to add agents in Copilot Studio is now in early access. The company expects a public preview to be available later in the year.

Copilot extensions and connectors

Microsoft is also announcing the ability to add your own extensions to Copilot for your own data and operations. These allow you to build customizations and data access tools to suit your business requirements. Extensions can be built in Copilot Studio or using the Teams Toolkit for Visual Studio.

Additionally, the company is introducing Copilot connectors, which are ready-to-use components to get you started building extensions.

More from Build 2024

Stay tuned to ZDNET for the latest up-to-the-minute coverage of Build 2024, including even more coverage of the Redmond-based company's announcements.

You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, and on YouTube at YouTube.com/DavidGewirtzTV.

Editorial standards