Radiologists face mounting pressures from high imaging volumes, documentation fatigue, and fragmented systems that slow reporting workflows. On November 24, 2025, at RSNA 2025, Microsoft announced Dragon Copilot for radiology—a new preview feature integrated with PowerScribe One—to address these bottlenecks using generative, multimodal, and agentic AI.
What Microsoft announced
Dragon Copilot, Microsoft’s AI clinical assistant launched earlier in 2025, now extends to radiologists as a companion to PowerScribe One, the cloud-based reporting platform trusted by thousands. Available in preview for PowerScribe One customers, it embeds AI directly into workflows without disrupting familiar tools.
Key features include prior report summarization into concise bullets, in-workflow chat with credible sources and patient context, report optimization for billing using third-party AI, and AI-drafted content from image models like CXRReportGen. It leverages Microsoft Foundry’s catalog of over 50 healthcare AI models, including premium versions for chest X-rays and image analysis.
Partners like Lunit for mammography insights and Zotec for revenue cycle management integrate as agents within Copilot, while broader ecosystem players (Merge by Merative, CitiusTech) enhance cloud imaging.
Why it matters
Radiology reporting lags due to manual prior reviews, research toggling, and quality checks, contributing to burnout and delays. Dragon Copilot unifies AI—speech recognition from Dragon Medical One, ambient capture from DAX, and multimodal models—into one secure, extensible workspace on Azure.
Official Microsoft sources emphasize no workflow disruption, with quotes from leaders like Sean Cleary, MD, at University of Rochester: “Cloud-based solutions that work seamlessly… with minimal disruptions.”
Impact and implications
This could cut report turnaround times, boost diagnostic accuracy, and reclaim hours for patient care. Early adopters like Mount Sinai (announced Nov 5, 2025) signal momentum. However, as a preview, success hinges on validation—local model testing, governance for safety, and regulatory compliance.
For practices, contact Microsoft for demos at RSNA booth #1311 or via account executives. As of November 2025, this positions Dragon Copilot as a leader in AI-driven radiology efficiency, per Microsoft Industry Blog and product pages.
| Feature | Benefit |
|---|---|
| Prior report summarization | Concise context for faster interpretation |
| AI chat | Reliable answers without app switching |
| Report optimization | Reduces claim denials |
| AI drafts | Accelerates draft-first reporting |
(Word count: 428)
