Microsoft Unveils 'Agentic Users' - AI Agents That Function as Independent Digital Employees

Microsoft teases agents that become 'independent users'

Microsoft has teased what it's calling "a new class" of AI agents "that operate as independent users within the enterprise workforce." "Each embodied agent has its own identity, dedicated access to organizational systems and applications, and the ability to collaborate with humans and other agents," states a Microsoft product roadmap document. "These agents can attend meetings, edit documents, communicate via email and chat, and perform tasks autonomously." Redmond will sell these "agentic users" in the "M365 Agent Store" and make them discoverable in its Teams collaborationware tools. Microsoft licensing specialist Rich Gibbons says he's seen additional documentation provided to M365 admins that mentions a license called "A365" - he thinks that's a reference to a product called "Agent 365" - and which states that "Admins assign the required A365 license at the time of approval. No additional Microsoft 365 or Teams license is required." A Microsoft MVP named João Ferreira appears to have seen the same document, and also mentioned the A365 license and shared a screenshot of an Agent 365 management page. Ferreira says agents will have their own email address, Teams account, an entry in enterprise directories (either Entra ID or Azure AD), and even a place on the org chart. "They can participate in meetings, send and receive emails and chats, access enterprise data, and learn from interactions to improve over time," he wrote. Microsoft's documents suggest the agents will debut later in November. As the software giant's annual "Ignite" conference kicks off next week, The Register fancies they may be one of the things Microsoft announces at the event. Ferreira said Microsoft will stage a "targeted release" of A365. Whenever it announces this offering, Redmond will doubtless make breathlessly optimistic pronouncements about their ability to improve productivity and profitability. Gibbons has looked past that and tried to guess what they'll cost to run. "Microsoft recently launched the Copilot Credit Pre-Purchase Plan (P3)," he wrote, noting that its base tier offers customers the chance to buy 300,000 credits. He says the software behemoth is "moving more and more to a consumption based pricing model, which is inherently much harder to forecast for customer organisations. Here, where we're going to have AI agents doings things off their own back - how are you supposed to predict usage/consumption in those scenarios?!" "As well as the licensing and cost concerns, I am also wondering how an organization manages these agents," he added. "If they can join meeting and send emails/messages to people - what happens if they go rogue? It could be sending sensitive data to the wrong people, providing incorrect information, or it could be sending strange or offensive messages...how is that to be prevented, monitored, and acted upon?"

Microsoft is working on a "new class" of AI agents that could change everything in your workforce

The agents are even described as having their "own identity" - but how will they work? Microsoft has revealed some early details of a "new class of AI" agents which it sees being able to enhance existing workforces by taking over some of the most repetitive administrative tasks. The upcoming "Agentic Users" will be just that - users - with their "own identity" and "dedicated access to organizational systems and applications." Microsoft also promises Agentic Users will be able to "collaborate with humans and other agents," diversifying the workforce beyond just humans. Tracked on the Microsoft 365 roadmap under the title "Microsoft Teams: Discovery and creation of Agentic Users from Teams and M365 Agent Store," the in-development update alludes to a November 2025 rollout. "These agents can attend meetings, edit documents, communicate via email and chat, and perform tasks autonomously," Redmond added. Maybe, then, Salesforce CEO Marc Benioff's recently-declared vision for the future has legs after all - having said today's CEOs will be the last to only manage humans - all future generations will also be tasked with managing AI agents as part of their workforce. Microsoft licensing specialist Rich Gibbons suggests these Agentic Users could have their own identity within organization directories via Entra ID or Azure ID - even their own email addresses and Teams accounts. Gibbons also notes that separate Agent 365 licenses could be required, though it's unclear how these will coexist with Copilot credit usage. Maybe Copilot will be distinguished as a human aid, while Agentic Users become their own whole separate category. With Microsoft Ignite just around the corner on November 18-21 2025, it's possible that Agentic Users could launch then, but for now, we're left only imagining how this new type of hybrid workplace could look.

Microsoft's new AI agents could soon join your office as digital employees

What's happened? Microsoft is working on a new kind of AI agent that doesn't just assist you, but actually acts like a full-fledged employee. These so-called "Agentic Users" will soon have their own email, Teams account, and company ID, just like a human coworker. These agents will be sold in "M365 Agent Store" and likely tied to a new license (reportedly called "A365" or Agent 365) rather than existing under the standard Microsoft 365 or Teams licensing. The rollout could start later this November, according to internal documents spotted by The Register. This is important because: These AI agents aren't like Microsoft Copilot or Claude assistant because they'll be able to send emails, join meetings, edit files, and complete tasks autonomously. In other words, Microsoft is building digital employees. It could dramatically change workforce dynamics as humans may collaborate with, manage, or even supervise AI employees. Productivity could increase, but so could complexity around management, security, and ethics. Recommended Videos Why should I care? If you work in an office, your next teammate might not be human. AI employees could take over routine tasks like scheduling, reporting, or customer responses, leaving people to handle the creative and decision-making parts. But it also raises tricky questions: Who supervises an AI that can act independently? What happens if it sends the wrong email, or worse, goes rogue? How do you keep track of software that behaves like a person? OK, what's next? Keep an eye on Microsoft Ignite 2025 (Nov 18-21) as there's a chance Microsoft could use the event to pull back the curtain on its new generation of AI agents. Once that happens, it'll be worth watching how companies actually put them to work and what new rules emerge around them. See where they show up first: admin work, customer support, or coding tasks could be early test zones. Watch the pricing model: will Agent 365 charge per agent, or bundle with existing Microsoft 365 plans? Expect new HR, legal, and compliance laws once AI agents start acting like real employees.

AI workers are getting their own email, ID and Teams accounts

If you work in an office, your next coworker might not be human at all. Workers are already well-acquainted with artificial intelligence in the office, using AI tools to take notes, automate tasks, and assist with workflow. Now, Microsoft is working on a new kind of AI agent that doesn't just assist, but acts as an employee. These "Agentic Users" will soon have their own email, Teams account, and company ID, just like a regular coworker. "Each embodied agent has its own identity, dedicated access to organizational systems and applications, and the ability to collaborate with humans and other agents," states a Microsoft product roadmap document. "These agents can attend meetings, edit documents, communicate via email and chat, and perform tasks autonomously. The rise of AI has already spelled death for middle management, and is having a "significant and disproportionate impact on entry-level workers in the American labor market," according to economists at Stanford's Digital Economy Lab.

Microsoft's Future AI Agents Will Behave as Independent Users

These agents can attend meetings and edit documents autonomously Microsoft is developing a new class of artificial intelligence (AI) agents for enterprises that can operate as independent users. The Redmond-based tech giant has listed the upcoming tool in its product roadmap for Microsoft 365. The company says, unlike typical agents that are meant for a specific purpose, these will be general-purpose embodied AI, complete with a unique digital identity, that can perform various tasks autonomously. The company has touted a release date of November; however, it is unclear when these agents will be released for its enterprise users. Microsoft Is Building General-Purpose AI Agents In its Microsoft 365 Roadmap page, the tech giant has now added a new product dubbed Agentic Users. These are currently under development with a tentative rollout estimated for November. The company says these AI agents will be available in the M365 Agent Store and Teams. A description of the product calls it "a new class of AI that operates as independent users within the enterprise workforce." What that means is that these AI agents will not be part of a particular system or operation, like most agentic tools, and instead, they can perform a wide range of tasks independently. To allow them to move from one software and platform to another, the Agentic Users will feature an identity and dedicated access to organisational systems and applications, Microsoft said. Coming to what these agents can do, the company says these tools can attend meetings, edit documents, communicate via email and chat, and perform other tasks autonomously. These agents will also have the ability to collaborate with humans as well as other agents. Rich Gibbons, a Microsoft licensing specialist (via The Register), mentioned in a blog post that he has viewed documentation provided to M365 admins. The documents are said to call the Agentic Users A365 and highlight that admins will assign the required license at the time of approval, and no additional license will be required. "So there will be at least one A365 license, but the word 'required' suggests, to me at least, that we will see a range of A365 licenses for different 'roles'," he added. Another admin, João Ferreira, saw the same documents and claimed that these AI agents will have their own email address, Teams account, entry in enterprise directories, and a place in the org chart.