This post is part of our “AI Agents in the Enterprise” series. Explore how IT, business, sales, and executive leaders can harness AI agents to transform their organizations. In this post, we focus on how you can empower your IT to achieve more by using AI and Microsoft Power Platform. We wrote this article based on keynote sessions by Microsoft at the European Power Platform Conference 2025.
In the age of AI, IT leaders are no longer just system stewards, they are architects of transformation. As organizations shift from using AI copilots to deploying autonomous agents, the role of IT becomes even more strategic. Microsoft Power Platform, with its evolving capabilities in Copilot Studio, Dataverse, and agent orchestration, is at the heart of this shift.
This blogpost explores how IT can lead the charge in building scalable, secure, and intelligent agent ecosystems that redefine how work gets done.
The shift from copilots to agents
AI copilots have already become indispensable in many organizations, helping users summarize emails, generate content, or automate simple workflows. But copilots are inherently reactive, they wait for a prompt. Agents, on the other hand, are proactive. They can reason, act autonomously, and collaborate with other agents and humans to drive business outcomes.
As Ryan Cunningham (Corporate Vice President, Power Platform Intelligent Applications at Microsoft) put it during the European Power Platform Conference, “We’re moving from augmenting individuals with copilots to empowering teams with digital coworkers, agents that can take initiative, make decisions, and orchestrate complex processes”.
Why IT is central to the agentic era
Deploying agents at scale isn’t just about building smarter bots. It requires a robust foundation of governance, security, data integration, and lifecycle management—areas where IT already excels.
Here’s how IT can lead:
- Architecting the agent platform
Power Platform is rapidly evolving into a full-fledged agent platform. IT teams can use Copilot Studio to define agent instructions, connect to enterprise data via Dataverse, and orchestrate workflows using agent flows and triggers.
With support for the Model Context Protocol (MCP), a standard that allows agents to interact with external tools and systems, APIs, and even other agents, you can enable multi-agent collaboration across departments. - Managing governance and security
Autonomous agents must operate within clearly defined boundaries. IT can enforce these through:
- Environment groups and deployment pipelines
- Role-based access control and data loss prevention (DLP) policies
- Agent sharing limits and approval workflows
- Enabling extensibility and customization
Every organization has unique data, processes, and systems. IT can extend agents using:
- Custom connectors to legacy or proprietary systems
- Azure AI Foundry models for domain-specific intelligence
- Enhanced connector SDKs to expose internal tools as agent-accessible services
Real-world examples: agents in action
Microsoft’s internal support agents
Microsoft deployed a team of agents to handle customer support cases. These agents autonomously triage issues, suggest solutions, and escalate only when necessary—reducing resolution time and improving customer satisfaction.
Another example: supplier discovery agent
Microsoft also showcased another example in which a company built an agent that identifies alternative suppliers during disruptions. It uses custom prompts, risk scoring, and agent-to-agent orchestration to resolve issues in minutes instead of hours.
How to get started: crawl, walk, run
IT leaders can guide their organizations through a phased approach:
- Crawl: Start by using copilots in Microsoft 365 and Power Platform. Encourage experimentation.
- Walk: Build simple agents for repetitive tasks (e.g., data entry, lead qualification).
- Run: Deploy multi-agent systems for complex workflows. Integrate with enterprise systems. Measure impact.
As Leon Welicki (Vice President of Product & CPO, Power Apps at Microsoft) noted, “You don’t need to wait until 2028 to start. Almost everything we showed today is available now”.
Final thoughts: IT as the agent boss
In this new era, IT isn’t just enabling automation, it’s orchestrating intelligence. By embracing Power Platform’s agent capabilities, IT leaders can:
- Drive exponential productivity gains
- Reduce operational costs
- Improve data quality and decision-making
- Future-proof their organizations
The tools are ready. The platform is mature. The opportunity is now. So the question is: will your IT team be the boss of your agents, or be left managing the chaos they create?
As we’ll explore in later posts, agents can also transform sales and executive decision-making.
Next in the series: 10X Your Business Outcomes: How AI Agents Are Redefining Operational Efficiency.