AI copilots and agents
A key area of AI development discussed by Satya Nadella, representing a shift from simple chat interfaces to autonomous agents that can perform actions and manage tasks for knowledge workers.
First Mentioned
1/22/2026, 4:20:09 AM
Last Updated
1/22/2026, 4:21:25 AM
Research Retrieved
1/22/2026, 4:21:25 AM
Summary
AI copilots and agents represent a significant evolution in artificial intelligence, transitioning from simple chat-based assistants to autonomous "agentic" systems capable of independent decision-making in complex environments. Microsoft CEO Satya Nadella envisions these agents as "infinite minds" for knowledge workers, facilitating a business revolution where companies like Microsoft, Alphabet, and Meta achieve growth without increasing headcount. While "copilots" serve as collaborative interfaces providing real-time support, "agents" are specialized, goal-oriented tools that automate specific business processes—such as sales, finance, and service—with minimal human oversight. The development of these systems involves a robust AI stack, including "token factories" on Azure, and a shift toward hybrid AI that utilizes local hardware like GPUs and NPUs on Windows PCs.
Referenced in 1 Document
Research Data
Extracted Attributes
Concept
Infinite minds for knowledge workers
Infrastructure
Token factories and app servers on Azure
Adoption Strategy
Dual process of top-down strategic initiatives and bottom-up transformation
Operational Model
Autonomous decision-making without continuous human oversight
Hardware Requirements
GPUs and NPUs for local model execution
Technical Classification
Agentic AI or Compound AI systems
Timeline
- Satya Nadella discusses the evolution of AI agents as 'infinite minds' during a fireside chat at the World Economic Forum in Davos, Switzerland. (Source: Document 4e50eb82-56c2-4d20-910f-9a43912c1cd7)
2024-01-17
- Microsoft introduces the Agent Builder experience within Microsoft 365 Copilot, allowing users to create specialized agents grounded in organizational data. (Source: Web Search Results)
2024-05-21
Wikipedia
View on WikipediaAI agent
In the context of generative artificial intelligence, AI agents (also referred to as compound AI systems or agentic AI) are a class of intelligent agents distinguished by their ability to operate autonomously in complex environments. Agentic AI tools prioritize decision-making over content creation and do not require human prompts or continuous oversight.
Web Search Results
- Copilot and AI Agents
A copilot, such as Microsoft 365 Copilot, is an AI-powered assistant that can help you be more productive and creative by providing real-time support, suggestions, and contextual guidance. Agents are specialized and can be used with a copilot to perform specific tasks, often with minimal input from the people that use them. They can respond to and resolve user inquiries in real time, or they can operate independently, taking specific actions based on data and predefined goals. They can also run business processes, adapt to new challenges, and improve over time. [...] If agents are like apps on an AI-powered interface, then a copilot is the interface that allows you to interact with these agents. Microsoft 365 Copilot, for instance, features a constellation of agents, including Microsoft 365 Copilot for Sales, Microsoft 365 Copilot for Service, and Microsoft 365 Copilot for Finance, to help you get things done. [...] ## Key takeaways Get an overview of the relationship between a copilot and AI agents. Discover the capabilities of AI agents, including task automation, data analysis, decision-making, and adaptability. Understand the different types of AI agents—and when to use them. Dive into the technology that gives AI agents the ability to communicate, learn, and adapt. See examples of AI agents in action. Get guidance on how to implement AI into your workflows or systems. Learn how AI agents are transforming business operations. Find out what’s in store for the future for AI agents. ## How a copilot relates to AI agents
- What is an AI Copilot? How is It Different From an AI Agent?
## What is an AI Agent? AI agents are designed to take on tasks independently while keeping you in the loop. Unlike copilots, which assist and collaborate with you, AI agents work autonomously to complement your efforts and free up your time for more strategic work. These goal-oriented systems are built to achieve specific outcomes, whether it's managing workflows, processing data, or responding to customer queries. They operate based on predefined rules—set parameters or guidelines established by you or your organization—ensuring that the actions they take are aligned with your goals. [...] ## AI Copilots vs. AI Agents: Key differences AI copilots vs. AI agents represent two distinct types of artificial intelligence, each designed to support specific needs in your workflow. While both enhance efficiency, they differ significantly in purpose, interaction style, and reliance on human involvement. Here’s a closer look at how each can work for you: [...] AI Agents: Once set up, AI agents operate autonomously, handling tasks based on predefined rules or algorithms. The need for human-in-the-loop is minimal—initial configuration is all that’s required. For example, an AI agent tasked with managing inventory will automatically reorder stock when thresholds are met, ensuring efficiency without intervention. This minimal dependence on human input makes AI agents perfect for repetitive tasks where speed and accuracy are critical.
- Genesys Agent Copilot
## Start-to-finish AI-powered support in a single solution Many businesses are only scratching the surface of AI’s potential to augment your human staff and increase efficiency — but Genesys Cloud Copilot goes farther. This AI solution is alongside agents and supervisors alike at all stages of their work, driving efficiency and elevating them to focus on high-value tasks. ### Deliver actionable advice and AI-enabled assistance in one solution Our AI copilots provide contextual insights for supervisors and concrete actions agents should take next. ### Make insights and analytics more accessible Surface insights on customer interactions at the time of need, so agents and supervisors have easy access to information. ### Drive continuous optimization of customer experiences [...] ### What is an AI agent copilot? AI agent copilot, sometimes called agent assist, refers to AI technologies that support human agents in handling customer interactions. These automated systems can suggest responses, provide relevant information in real time to answer questions before they’re asked, automate and execute business processes and tasks, and help agents resolve customer queries more effectively and efficiently. ### What kind of AI does agent copilot use? [...] ### How does copilot’s AI functionality help supervisors deliver better coaching, improve agent engagement and enhance overall customer experiences? Insights generated from interactions pinpoint areas of improvement, allowing supervisors to coach agents based on personal development needs. Personalized career development and coaching engages agents because they feel like an individual. Refined coaching helps agents improve skills that improve the overall customer experience and drive consistency. Coaching and performance can be tracked through real-time AI dashboards in performance management.
- Build Your Own AI Agents with Microsoft 365 Copilot
141 comments ### Transcript: I'm going to take you through the agent Builder experience which is part of your Microsoft 365 co-pilot license and it allows any user no skill required to start building their own agents let's just level set on a couple of the concepts here so that you know what we're doing and why we're doing it this is co-pilot Microsoft 365 co-pilot I'm in the work experience here and this is my starting point for all of my AI experiences but what if I want to tweak it a bit customize it a bit have it work for just a specific business process flow or to automate a specific thing that I do at work that's where Agents come in and what I'm showing you now is sort of the first early release of this there's plenty more coming please subscribe to the [...] building agents these ones are at the moment quite basic but once actions come in here there'll be a little bit more sophistication and capability for things that you can do but the other thing that's been announced is this is the main co-pilot studio now this one if you want to do it this way you do need a separate license for co-pilot Studio but here you've got a lot more functionality you can start to build out specific topics the access to actions is completely there there's a lot more knowledge sources you can work with than just the the SharePoint and toggling on that website knowledge source and you can build one of these agents with all of that richness involved and then bring it back inside that Microsoft 365 co-pilot in that same user experience way that [...] we'll let it go while I keep talking here your Microsoft 365 co-pilot work experience by default unless your admin has switched this capability off which they can do is grounded in all of your documents and also grounded in the web so what I'm saying now is I'm basically creating an agent here which is scoped right down I'm saying I want an agent that is just looking at that particular folder in SharePoint and nothing else not going to look at my emails or my uh my teams chats it's not going to be grounded in the web because I switched that off so this is one example of something you might do with it just to create a really refined scope agent please look at this you can imagine if you had a folder of documents about a particular project that a team was working on you could
- Get started with agents in Microsoft 365 Copilot
Applies To Microsoft365.com Agents extend what you can do with Microsoft 365 Copilot—you can customize your Copilot experience by connecting to your organization’s knowledge and data sources, as well as by automating and executing business processes. These AI-driven agents can perform a variety of tasks, working alongside you to offer suggestions, automate repetitive tasks, and provide insights to help you make informed decisions. You can use pre-built agents, agents created by others in your organization, or create your own. You can also share a link to an agent if you want to help someone find and use a specific agent. Agents are available in both the Web and Work tabs. ## Benefits of using agents