新たなMicrosoft AIエージェントが企業の近代化を支援

2 Min ReadMicrosoft has introduced new AI agents designed to help enterprises digitally transform with maximum impact and minimum fuss.The releases should help organizations by bringing IT departments and developers into a single workflow, reducing the complexity of modernization, according to a blog post post by Jeremy Winter, Microsoft’s Azure platform chief product officer.The Azure Copilot Migration Agent, introduced last week, is now in public preview. It is designed to help organizations embed AI across discovery, assessment, planning and deployment.The agent turns “migration from a one-time project into a continuous modernization motion” and can be used on most mission-critical servers, applications and databases, Winter said. “The agent uses what customers already know about their environments to create this continuous modernization system providing early clarity on their inventory … how it all works, what it costs, and what’s worth modernizing.”Related:OpenAI Updates Agents SDK, Aims at Building Secure AgentsThe result, he added, is that tasks that previously took months of manual planning can now be executed in minutes, thanks to a brief conversation with the agent to generate a data-driven plan.Meanwhile, GitHub Copilot enables developers to use agents that can transform legacy applications into cloud- and AI-ready systems. Microsoft also rolled out a new capability for GitHub Copilot, the modernization agent, now in public preview.This agent essentially acts as an orchestrator, simultaneously running multiple code assessments, creating unique modernization plans for each application, and executing them with automated framework and runtime upgrades, according to Microsoft.Microsoft said Azure Copilot and GitHub CoPilot working together will address the disconnect between code and infrastructure, within which developers have modernized code without knowing how it would run, scale, or be governed in the target environment.“We’re moving away from one-off use of agentic tools to consistent, repeatable agentic execution," Winter said in the blog. This is the importance of pairing agentic tools with structured delivery models, so organizations can scale modernization confidently across teams and environments.”Separately, releases a public preview of Fireworks AI on Microsoft Foundry, the vendor's unified platform for developers and enterprises to build, deploy, optimize and manage AI applications. It also gives developers access to inferencing for open models, including optimized deployments for custom weight models.Among the models available through Fireworks AI on Foundry are DeepSeek V3.2, OpenAI gpt-oss-120b and Kimi K2.5 from Moonshot AI.Related:As AI Infosec Woes Heighten, IBM Intros Autonomous Security ServiceAbout the AuthorContributing WriterGraham Hope has worked in automotive journalism in the U.K. for 26 years, including spells as editor of leading consumer news website and weekly Auto Express and respected buying guide CarBuyer.