That’s the promise of Adaptive Intelligence, and it’s now within reach. In our recent Ignite session, Jason Huang and I walked through exactly how to build enterprise-grade agents that don’t just automate—they think, adapt, and elevate work. See this in action:

Why Adaptive Intelligence Matters

Security and governance are table stakes. The real question is: are your agents smart enough to be useful?

The real differentiator is intelligence. Agents that understand your business. That remember how you work. That act proactively, not reactively.

That’s Adaptive Intelligence. And it’s built on two foundations:

1. Work IQ + Dataverse: Work IQ is the intelligence layer that gives agents context, memory, and inference. Dataverse extends Work IQ with trusted business data: giving agents the ability to reason over your structured enterprise data while staying inside governed boundaries. Together, they transform Dataverse from a system of record into an intelligent system of action.
2. Agent 365: the control plane that makes agents enterprise-ready by default. It handles registry, access control, visualization, interoperability, and security—so IT doesn’t lose control as agents scale.

When Work IQ, Dataverse, and Agent 365 come together, agents don’t just automate. They think. They adapt. They elevate work.

How We Built It

The agent in our demo, Aligna, is an AI teammate for a fictitious construction company (Zava Construction). In the demo, a site coordinator chats with Aligna in Teams: “@Aligna, I’m at the Adatum site, log this plumbing issue, inform the client…” — and Aligna handles the rest.

Here’s how each component powers that workflow:

Step 1: Microsoft Agent Framework

Aligna is built on the Microsoft Agent Framework—it’s what enables the agent to receive the Teams prompt, decide which tools to call, and orchestrate the entire workflow. When the user describes the plumbing issue, the Agent Framework routes the request to the right tools (Dataverse MCP for logging, Mail MCP for communication) and manages the execution flow.

Step 2: Dataverse MCP

When the user says “log this plumbing issue,” Aligna uses Dataverse MCP (now in GA) to create the issue record directly in Dataverse—no forms, no app-switching. All business data (clients, contacts, projects, vendors) lives in Dataverse, and MCP provides the standardized interface for Aligna to read and write that data through natural language.

Step 3: Mail MCP Server

When the user says “inform the client,” Aligna uses the Mail MCP Server to draft and send the email—all within the same conversational flow. The client gets notified about the issue without the user ever leaving Teams.

Step 4: Agent 365 SDK

Every action Aligna takes—logging the issue, sending the email—is recorded via the Agent 365 SDK for compliance and audit. IT gets full observability into what the agent did, when, and why—without requiring extra instrumentation from developers.

Step 5: Dataverse SDK for Python

For developers building agents like Aligna, the Dataverse SDK for Python (in Public Preview) provides direct programmatic access to Dataverse—ideal for custom data operations, schema setup, batch processing, or integrating with Python-based ML pipelines. Also leveraged: Azure OpenAI for model inference

Turn Your Business Data into Agent Intelligence

The tools are here. The documentation is live. Now it’s your turn.

Start building agents that truly understand your business:

1. Connect to your data — Use Dataverse MCP to give your agents instant access to business data. It’s GA and ready for production.
2. Write in your language — Python developer? The Dataverse SDK for Python lets you prototype in minutes. .NET team? The Agent Framework has you covered.
3. Make it enterprise-ready — Register with Agent 365 and your agent inherits identity, governance, and observability from day one.

Your competitors aren’t waiting. The agents your organization needs—agents that think, adapt, and elevate work—you can build them now.

Read More

Related Documentation:

• Dataverse MCP: Connect to Dataverse with MCP 
• Dataverse SDK for Python: SDK Overview
• Agent 365 Overview: Agent 365 Documentation
• Agent 365 SDK: Agent 365 SDK Overview
• Microsoft Agent Framework: Agent Framework Documentation
• Mail MCP Server: Mail MCP Server Reference

Related Ignite & Blog Resources:

• Ignite 2025 Book of News
• Microsoft Agent 365 Interoperability for Smart, Secure Productivity – Microsoft 365 Developer Blog

The post Build Agent 365-ready Adaptive Intelligence with Work IQ and Dataverse appeared first on Microsoft Power Platform Blog.

Unlock Deeper Insights with Multiline Text Columns and File Columns in Dataverse Knowledge 

Agents built in Copilot Studio using Dataverse Knowledge can now intelligently search across multiline columns and file columns, making it easier than ever to find the information you need. Just ask in natural language, and these agents will surface the most relevant results—no matter where the data lives. Searching through long comments or attached files is now easier than ever with Dataverse Knowledge.  

Here are some examples you can try out: 

• “Which phones have the best reviews for creators?” 
• “Show me the phone that’s a balance for storage and battery life” 
• “What are some reviews of Contoso phone?” 

and you’ll get the exact results you need—even if that information is buried inside a file or multiline text. 

Note: Dataverse Knowledge in Copilot Studio does not support images and tables in files and will only return results in the language of the content. For example, if a document is written in English, it must be searched using English terms. Multi-lingual/image/table search from within files will be included in a future update. 

Improved Answer Consistency with Dataverse Knowledge 

We know it can be frustrating to get different search results for the same question, especially when you rely on accurate information for your work. That’s why we’ve added a new behind-the-scenes update that makes sure your agents grounded in Dataverse Knowledge can produce the same reliable answer every time you repeat a search—even if you ask the same question at a different time. It’s perfect for common work tasks like checking support issues, reviewing contracts, or pulling reports. Consistency and trust in your results are now built right in! 

For example, a query such as – “find contracts expiring within the next 60 days” in compliance or legal review scenarios to demonstrate how accuracy and reliability are maintained for repeated business-critical queries across time. 

Note: This update requires no set up and should be available out of the box. 

Start exploring Dataverse Knowledge with Smarter Search

With support for multi-line and file columns, improved answer consistency, and reliable caching, Dataverse Knowledge in Microsoft Copilot Studio is now more powerful than ever. These updates remove long-standing limitations and make it easier to surface business-critical insights—no matter where they’re hidden. 

Whether you’re building agents to handle support tickets, review contracts, or analyze feedback, Copilot Studio now gives you the tools to deliver faster, more accurate, and more consistent results:

• Try natural language queries on long notes and files 
• Experience consistent results with built-in caching  

See resources to learn more and get started:

• Read this quick set up guide: Support for multiline text and file columns 
• Read documentation: How to upload files into a file column in Dataverse 
• Watch a demo: Unlock your Dataverse data with secure and powerful search 

The post Dataverse Knowledge in Copilot Studio: Multiline Text Columns and File Columns. appeared first on Microsoft Power Platform Blog.

What Is the Model Context Protocol (MCP)?

MCP is an open protocol that bridges large language model (LLM) applications with external data sources and tools. It acts as a bridge between AI assistants and enterprise data, enabling secure, standardized access to Dataverse resources. When Microsoft Dataverse acts as an MCP server, it enables intelligent, secure, and standardized access to your data—tables, records, and more—across a variety of MCP clients.

Why It Matters

Dataverse MCP Server isn’t just another connector—it’s a paradigm shift. Here’s why:

• Unified AI-to-Data Access: MCP standardizes how LLMs like Copilot Studio agents, Claude, and GitHub Copilot interact with Dataverse. This means that your enterprise data is now fully extensible and platform agnostic.
• Natural Language interface: Users can ask questions like “show me my contacts” or “apply 10% discount to Widget123 product” and get real-time response from your Dataverse environment. Under the hood, MCP clients translate these natural language requests and execute one or more built-in tools in Dataverse.
• Powerful Tooling: Dataverse MCP Tools currently support built-in tools that allow users to insert or update data in tables, view tables and their description, read data from a table, search knowledge, view and execute prompts available in environment. See List of tools available in Dataverse MCP Server documentation for more details.
• No SDKs, no custom APIs—just natural language and secure execution.

MCP Clients are capable of determining the appropriate tools to use – so while the below image illustrates a workflow that an agent may take, the tools used by the MCP Client may vary from time to time.

Works Across Your Favorite MCP Hosts and Clients

Microsoft Copilot Studio

Connecting is as simple as adding the Dataverse MCP Server as a tool to your agent. No local setup required. You can immediately test commands like: “List tables in Dataverse”, “Describe table account”.

Using Dataverse MCP in Copilot studio is a compelling alternative to having agent makers using custom Dataverse connectors to manually create flows to store, update, or retrieve the data. Now, you as a maker, can simply add Dataverse MCP from the Tools tab inside an agent in Copilot Studio. Under the hood, when an agent receives user messages such as “What is the status of my support ticket #12345”, this translates to the agent calling the appropriate Dataverse MCP Tools such as determining the appropriate table to read via list_tables or describe_table, then read_query against that table.

Now makers can build dynamic, intelligent workflows with conversational agents that talk directly to Dataverse—no flows, no APIs, just prompts + tools!

Claude Desktop

With a quick configuration update, Claude can tap into your Dataverse environment. You can even toggle which tools are available per MCP server, giving you full control.

Business users often has to rely on their IT team  to create dashboards. Now, with Dataverse MCP in Claude, the workflow is much streamlined. Here’s an example to illustrate it.

• User can request in Claude: “Create a dashboard showing opportunities by region”
• This translates to Claude identifying which table(s) opportunities are stored using list_tables or describe_tables tools, then using read_query to get the information for Claude then to perform a code-gen to render an interactive dashboard.

Dataverse MCP, thus, empowers business users to explore and interact with their business data stored in Dataverse conversationally—zero-code access to real-time, live insights and automation.

GitHub Copilot in VS Code

Reuse your Claude configuration or set up a new one. Once connected, you can interact with Dataverse directly from VS Code GitHub Copilot in Agent mode. For example: “store the attached csv file as Employee in Dataverse”. Under the hood, GitHub Copilot will automatically inspect the data and upload the data into the table – all achieved by invoking tools like list_tables, describe_table, and create_record tools to achieve this.  

With Dataverse MCP, developers see improved productivity as they don’t need to open multiple portals or tools. Schema inspection, CRUD actions, and code generation happen inside the IDE!

⚠️ A Note on Preview Features

This is a preview capability—perfect for early adopters and innovators. While not yet production-ready, it’s a fantastic opportunity to explore, experiment, and provide feedback to shape the future of AI + data integration.

Coming Soon: Remote MCP Server + Full Table Operations!

We’re thrilled to announce that we’re bringing even more power and flexibility to Dataverse MCP! Starting end of July, we’re rolling out a Private Preview of two major capabilities:

• Remote MCP Server – drastically simplify agent setup by just updating the environment URL, no complex configurations needed
• Expanded Tooling – including long-awaited operations like create_table, update_table, and delete_table, enabling full control over your Dataverse schema through agents

If you’re as excited as we are and want early access, Click here to express your interest and be among the first to shape the future of agent-powered data platforms.

Final Thoughts

The Dataverse MCP Server represents a strategic evolution in how makers, developers, and data professionals interact with enterprise data. By enabling natural language access to Dataverse, it simplifies development and accelerates innovation across the Power Platform ecosystem and beyond.

Whether you’re designing data models, querying records, or building intelligent agents, MCP makes Dataverse the brain behind your AI.

Learn more:

• Dataverse MCP – Part 1: Starting in GitHub Copilot – YouTube
• Connect to Dataverse with model context protocol (MCP) – Power Apps | Microsoft Learn
• We invite you to explore the preview, share your feedback, and join the conversation using #DataverseMCP on LinkedIn.
• Submit ideas about Dataverse MCP using Dataverse as a category.

The post Dataverse MCP Server: A Game Changer for AI-Driven Workflows appeared first on Microsoft Power Platform Blog.

In every organization, data fuels critical business processes. Yet, much of this data is trapped in formats and workflows that make it difficult to access, understand, and use effectively. A common example is invoices arriving as PDF attachments via email—disconnected from the broader business data ecosystem. These files are hard to search, lack structured linkage to financial systems, and often require tedious manual intervention to process. 

For technical decision-makers, this isn’t just an operational headache—it’s a strategic liability. 

• Manual data processing drains valuable resources, pulling skilled teams away from higher-value work. 

• Data integrity suffers, increasing the risk of errors, missed payments, or compliance issues. 

• Lack of discoverability means critical insights remain hidden, slowing down decision-making and reducing the agility of the business. 

In a data-driven world, relying on fragmented and manual workflows undermines the ability to scale, innovate, and stay competitive. The challenge is not having the data—it’s unlocking its potential efficiently and at scale. 

This post will explore how, with Microsoft Copilot Studio and Microsoft Dataverse, businesses can build intelligent agents that automatically turn fragmented data into structured, actionable insight. These AI-powered agents extract, validate, and enrich business data as it arrives, store it in Dataverse for reuse, and loop in humans only when needed. The result? Cleaner data, faster decisions, and a foundation for scalable, intelligent automation. 

Learn how Dataverse provides a secure, scalable, and agent-ready foundation for your data strategy in the Why Dataverse Overview. 

Figure 1: Built for Data Agents overview, with Dataverse as the underlining data platform for Agents. 

A Better Way: Activating Data with Autonomous Agents (with Dataverse at the Core) 

Modern businesses now have access to a more scalable, intelligent approach to data processing: Data Agent architectures powered by autonomous agents with human-in-the-loop oversight. 

What does this mean in practice? 

• Autonomous agents handle repetitive data processing tasks intelligently: extracting, validating, and enriching data from unstructured sources such as emails, PDFs, tickets, and more. 

• Dataverse serves as the unified data platform: offering a secure and consistent way to store business data—structured or unstructured—so it can be easily discovered, related, and used across business applications and workflows. 

• Humans stay in the loop for oversight and decision-making: agents escalate exceptions, low-confidence matches, or contextual queries to human reviewers, who validate and finalize the data. 

The result: Intelligent, high-quality data delivered faster and more reliably, and centralized in Dataverse for seamless integration across business apps, analytics, and automation workflows. 

This architecture directly aligns with how Microsoft Copilot Studio agent flows work: orchestrated automation enables conversational agents (or flows) to autonomously process and route data, while intelligently involving human judgment when context or precision is crucial. With Dataverse at the heart, every interaction, enrichment, and validation feeds into a scalable, governed data platform—ensuring data isn’t just processed but activated for business value. 

While autonomous agents in Copilot Studio can operate independently using built-in tools, makers can enhance their capabilities further with tools like the Model Context Protocol (MCP). MCP isn’t required to build autonomous agents, but it opens up new possibilities for agents to process or store data in Dataverse. This bidirectional flow ensures AI solutions remain context-rich, action-oriented, and aligned with Microsoft’s broader commitment to openness and extensibility. 

Through integration with MCP, agents gain powerful data handling capabilities such as: 

• Query: Discover table schemas and retrieve live Dataverse records via structured or natural-language queries; 

• Knowledge/Search: Lets agents “chat” about data – they can search tables and answer questions contextually without brittle hand-coded logic; 

• Upload (Create/Update): Insert new records or update existing ones in Dataverse, with built-in schema validation to maintain data integrity; 

• Generate (Grounded AI): Run custom AI prompts grounded in real data (e.g. summarize a record or evaluate sentiment). 

Because the MCP server honors Dataverse’s data model and access controls, agents can safely reason over and act on enterprise data. In practice, developers can simply add a Model Context Protocol tool in Copilot Studio and point it to their Dataverse. The agent then automatically queries and updates that data as part of its workflow – for example, posting a generated summary back into a record. 

Figure 2: Shows adding the Model Context Protocol Server tool within Copilot Studio. 

Learn more with: 

• Announcing new Microsoft Dataverse capabilities for multi-agent operations 

Get Started Building Intelligent Agents That Work for You 

As organizations move beyond basic automation, the next frontier isn’t just bots that complete tasks, it’s intelligent agents that understand, interact, and learn from business context. With Copilot Studio and Dataverse, you can create AI-powered agents that operate securely, respond to real-world triggers, and collaborate with humans to drive outcomes. 

We’ll walk through how to build and operationalize intelligent agents in six easy steps: 

Step 1: Trigger events from your business processes 

Agents can be triggered automatically based on business events—such as a new email, a service request, or a record added to a Dataverse table. These triggers, configured in Power Automate, initiate the agent’s workflow and response logic in Copilot Studio. 

Learn more with: 

• Add an event trigger 

Step 2: Equip agents with intelligent tools 

Once triggered, the agent uses Copilot Studio capabilities to process requests: 

• Knowledge Sources provide background information. 

• Topics guide the structure of the conversation. 

• Tools allow agents and MCP server(s) to perform tasks across systems. 

• Conversation History maintains context for continuity. 

• Instructions shape how the generative AI responds. 

Using these tools—along with Agent Flows and logic based on Model Context Protocol—agents can analyze data, perform multi-step actions, and interact meaningfully with users. 

Learn more with: 

• Agent flows overview 

• Use generative answers with knowledge sources 

• Create and manage topics 

• Add actions to custom agents 

Step 3: Use Dataverse as an intelligent data layer 

Dataverse brings structure and intelligence to your business data—so your agents can retain context, access relevant records, and contribute to a broader operational view. It gives Copilot agents secure, real-time access to the records, relationships, and history they need to work smarter across your apps and processes. 

With native support for the MCP, agents can draw from and contribute to Dataverse as part of a broader knowledge network. It’s a trusted foundation for building agents that are grounded in data, aligned with workflows, and ready to scale. 

Learn more with: 

• Add a Dataverse knowledge source 

• Use Dataverse table data as knowledge in Copilot agents 

Step 4: Enable autonomous agents with human oversight 

As agents become more capable, they also need to support human-in-the-loop scenarios. With conversation history and activity logs stored in Dataverse, business users can monitor agent performance, review outputs, and step in when escalation is needed—ensuring oversight, accountability, and trust. 

Learn more with: 

• Build an autonomous agent in Copilot Studio. 

• Unlocking autonomous agent capabilities with Copilot Studio. 

• Manage Dataverse auditing 

Step 5: Coordinate workflows with the Agent Orchestrator 

Agent Orchestrator enables the user to coordinate more complex multi-step workflows by managing how and when tasks are executed across autonomous agents. For complex scenarios, it collects inputs from initial triggers, stores intermediate responses for reuse, and drives execution using Copilot Studio actions, Power Automate flows, or custom APIs. With built-in logic for queuing, task routing, and fallback handling, Agent Orchestrator ensures that each step is carried out by the most suitable agent—making it easier to scale intelligent, resilient automations across the organization. 

Learn more with: 

• Orchestrate agent behavior with generative AI 

• FAQ for generative orchestration 

Step 6: Govern secure access and responsible usage 

All agent interactions are secured through enterprise-grade governance controls: 

• Role-Based Access Control (RBAC) ensures data access is scoped by role, table, and even row. Learn about RBAC. 

• Least-privilege access, audit logging, and connector risk policies support safe and responsible agent behaviour. Least privilege and connector governance in Power Platform. 

• Secure agent data access with an organized security framework that integrates platform technology, regulatory compliance, and administrative oversight—following best practices for managing access, reducing risk, and ensuring responsible agent behavior. Overview of role-based security in Dataverse. 

By combining Copilot Studio, Dataverse, and Power Platform tools, enterprises can create intelligent digital workers that adapt to the business, work alongside humans, and scale with confidence. This is how modern enterprises move beyond simple automation—and into a future powered by AI-first, data-smart systems. 

For more information watch Ryan Cunningham and Evan Lew discuss how to Build agent-first solutions with Power Platform and Copilot Studio. 

Figure 3: Workflow highlighting how Autonomous Agents interact with human and agent oversight, with Dataverse at the core. 

Real world application 

Velrada’s PowerRoster solution is helping frontline organizations reimagine workforce scheduling—reducing complexity and enabling smarter, faster decisions. At the heart of this transformation is ShiftLens, a new capability built with Copilot Studio that uses autonomous agents and intelligent workflows to automate manual scheduling tasks. ShiftLens is designed for high-pressure, high-variability environments such as healthcare and hospitality. When absenteeism is logged, ShiftLens records the key details into Microsoft Dataverse, updates the roster, and recommends staffing changes for manager approval. 

By eliminating the daily scramble to find last-minute replacements, team leaders and managers can focus more on delivery and less on logistics. This intelligent orchestration not only improves operational efficiency but also contributes to better staff satisfaction and customer outcomes. The demo below shows ShiftLens in action at a fictional care home—highlighting how AI-powered scheduling is making frontline workforce management more proactive, resilient, and human-centered. 

Figure 4: Video showcasing Velrada’s ShiftLens Intelligent Rescheduling product, powered by Microsoft Copilot Studio and Dataverse 

Conclusion 

The cost of manual processing isn’t just time—it’s missed insights, slower decisions, and operational drag. Copilot Studio and Dataverse offer a smarter path forward: intelligent agents that not only automate work but understand your data, collaborate with your teams, and evolve with your business. 

This is how leading organizations are moving from scattered workflows to coordinated, AI-powered systems. Whether you’re streamlining frontline operations or scaling data-driven decisions, now is the time to rethink what your business could do with agents that learn, act, and improve—built on a secure, governed foundation. 

Learn more about Copilot Studio and Dataverse

Explore the resources below to take the next step with data agent architecture powered by Copilot Studio and Dataverse. 

• Learn more about Data Agent Architecture  
• Explore how Microsoft is advancing intelligent agent design with Dataverse: Announcing new Microsoft Dataverse capabilities for multi-agent operations. 
• Learn how to Connect to Dataverse with Model Context Protocol. 
• Learn more about the Document Processor agent in Copilot Studio. 

From Microsoft Build 2025 

• Learn how GenAI for Enterprise: Intelligent Apps and Agents with Dataverse & MCP enables intelligent, grounded agents built on your enterprise data and systems. 
• Dataverse helps you turn business data into a platform for intelligent apps and automation with Dataverse for agents. 
• Copilot Studio revolutionizes document processing and content generation with Agents in action: Document processing 2.0. 
• Effectively leverage your enterprise Knowledge in Copilot Studio to enhance your agents with more relevant and contextual answers. 

The post Data Agent Architecture powered by Microsoft Dataverse  appeared first on Microsoft Power Platform Blog.