Microsoft Launches ‘Deep Research’ in Azure AI Foundry for Smarter Enterprise Automation

Microsoft is bringing powerful new AI research tools to Azure. The company yesterday launched the public preview of “Deep Research” under its Azure AI Foundry Agent Service.

This new offering lets developers build AI agents that can handle complex research tasks, sifting through web data, analyzing it in-depth, and generating transparent, traceable reports. Microsoft says it’s designed for enterprise users who want more than just chatbots. These agents can be embedded directly into apps and automated workflows.

Powered by OpenAI’s latest models, including GPT-4o and o3-deep-research, the system starts by clarifying the user’s query, then pulls recent web data using Bing Search grounding. From there, the AI synthesizes the information into structured, auditable outputs, citing sources and explaining its reasoning process along the way.

Microsoft is positioning this as a big step beyond tools like Copilot or ChatGPT. Deep Research isn’t just about summarizing the internet. It’s built for multi-step reasoning, real-time analysis, and full integration across business platforms.

That means you could, for example, run a research agent, feed the results into a slide generator, then automatically share the deck with stakeholders via Logic Apps. “Research is no longer a manual, one-off task, it’s a building block for digital transformation,” Microsoft says in a blog post.

The tool is available now in limited public preview, with enterprise-grade compliance, security, and pricing set at $10 per million input tokens and $40 per million output tokens. For enterprise teams aiming to automate deep knowledge work, this could be a game-changer.