How Microsoft’s AI Futurist uses Copilot to solve real enterprise problems

Microsoft’s Build conference this year delivered a unified message: AI agents are no longer experimental—they’re operational. The key to enterprise adoption lies in providing these agents with reliable context, governance, identity, and secure access to enterprise data. To make this vision a reality, Microsoft unveiled a new IQ framework designed to connect agents like GitHub Copilot, Microsoft Foundry, and Copilot Studio seamlessly across its ecosystem.

Among the most significant additions are Work IQ APIs, planned for release in mid-June, which will enable agents to interact natively with Microsoft 365 apps. Fabric IQ will extend this capability to structured business data, while Foundry IQ will unlock retrieval across enterprise knowledge and live web sources. A new Web IQ will serve as an agent-facing web search stack, eliminating the need for agents to navigate traditional interfaces. The company also introduced Scout, a personal work agent, and seven new in-house AI models under the MAI family, including MAI-Thinking-1.

At the heart of this strategy is Marco Casalaina, Microsoft’s VP of Products for Core AI and AI Futurist. His role blends hands-on experimentation with strategic foresight, ensuring the company stays ahead of emerging trends while addressing real-world enterprise challenges.

The AI Futurist: Bridging today’s tools with tomorrow’s possibilities

Casalaina’s position as AI Futurist isn’t just a title—it’s a function. He describes his role as the first person to test new AI innovations across Microsoft’s portfolio, from Foundry and Visual Studio to GitHub and Copilot. This hands-on approach allows him to identify gaps, evaluate usability, and refine products before they reach broader audiences.

"I get to see a lot of really cool stuff," Casalaina said in an interview ahead of Build. "Pretty much everybody sends me new things all the time—even today, I got something brand new just before this call." His approach aligns with what he calls "adjacent possiblist," focusing on the immediate future—what’s about to happen within the next year rather than decades-long predictions.

This perspective shapes how Microsoft designs its AI tools. Instead of prioritizing raw model performance alone, Casalaina emphasizes practical integration, governance, and scalability. "Enterprises need far more than just access to powerful models," he noted. "They need context, control, and security."

Model choice, customization, and the rise of hosted agents

At the foundation of Microsoft’s AI strategy is model choice. The company continues to integrate third-party models like Anthropic’s Claude while also expanding its in-house MAI family. These models are engineered for efficiency, enabling businesses to fine-tune them on proprietary datasets without sacrificing performance.

Recent additions include Claude Opus 4.8 on Azure and Foundry, alongside the new MAI models optimized for token efficiency and customization. This dual approach ensures enterprises aren’t locked into a single provider while still benefiting from Microsoft’s infrastructure.

Above the model layer, Microsoft is introducing hosted agents in Foundry, a managed environment where businesses can deploy, scale, and monitor agents without worrying about infrastructure. Foundry’s control plane adds another layer of governance, offering observability into costs, token usage, and agent correctness. Continuous evaluation tools help prevent drift, ensuring agents remain reliable over time.

The IQ framework: Unifying data, apps, and agents

The most transformative aspect of Microsoft’s agent strategy is the IQ framework—a set of interfaces designed to bridge the gap between agents and enterprise systems. Unlike traditional middleware, IQs act as agent-facing layers, abstracting away the complexity of interacting with disparate tools and data sources.

• Foundry IQ: Focuses on unstructured enterprise knowledge, enabling agents to retrieve and synthesize information without manual input.
• Fabric IQ: Bridges structured business data in Microsoft Fabric, Power BI, and related platforms, allowing agents to query datasets directly without navigating reports.
• Work IQ: Serves as the agentic interface for Microsoft’s productivity suite, including Outlook, Teams, Word, and SharePoint. Agents can automate workflows, extract insights, and interact with documents natively.
• Web IQ: Introduces a headless, agent-facing web search capability. Unlike traditional search engines, Web IQ performs autonomous browsing tasks, retrieves real-time data, and processes multimedia content efficiently.

Work IQ APIs, slated for release on June 16, will be the first to reach general availability, marking a critical step toward seamless agent integration across Microsoft’s ecosystem. The remaining IQs are designed to evolve alongside enterprise needs, ensuring agents remain adaptable as data landscapes shift.

What’s next for enterprise AI agents?

Microsoft’s latest announcements signal a shift from experimentation to execution. Enterprises are no longer asking if they should deploy AI agents—they’re focused on how to do it efficiently, securely, and at scale. The IQ framework and hosted agent capabilities address these concerns by providing the infrastructure, governance, and context required for real-world deployment.

Casalaina’s role as AI Futurist underscores a broader trend: the future of AI isn’t just about smarter models—it’s about smarter integration. By prioritizing usability, control, and interoperability, Microsoft is positioning itself as a leader in the next phase of enterprise AI adoption. As agents become more autonomous, the challenge will shift from building them to managing them—and IQs may well become the standard for how businesses interact with AI.