AI Scribes and Dictation Tools Move Deeper Into Radiology Workflow at St. Luke’s
St. Luke’s University Health Network is using PowerScribe One and Dragon Copilot to optimize radiology workflow. The deployment reflects a broader shift from experimental AI to workflow infrastructure that aims to reduce friction in routine clinical documentation.
Radiology AI is increasingly less about flashy detection and more about the unglamorous work of saving time. Documentation, report generation, and workflow orchestration may not sound revolutionary, but these are the areas where clinicians feel friction every day.
The appeal of tools like PowerScribe One and Dragon Copilot is that they target an obvious operational pain point: getting findings into the chart accurately and quickly. In a high-volume specialty, even small efficiency gains can compound into meaningful capacity improvements.
But workflow AI also creates a new kind of dependency. Once a health system builds processes around vendor-specific tooling, switching costs rise, and the quality of the integration can matter as much as the model itself.
This is a good example of where healthcare AI is headed: not replacing clinicians, but embedding itself into the operational layer of care. The winners may be the systems that can turn AI into durable infrastructure rather than one-off productivity pilots.