AI That Works With Radiologists, Not Against Them

Imaging volumes on the rise. Workforce not keeping pace. It’s an ongoing challenge, and for radiologists already dealing with demanding workloads, the solution isn’t simply more technology — it’s technology that makes their work easier. 

In a recent episode of the “Outcomes Rocket” podcast, Rad AI Co-Founder and President Jeff Chang, MD, discussed why the future of healthcare depends on understanding clinical workflows and building solutions that fit naturally into them. 

The Workflow Is the Starting Point

Dr. Chang noted that imaging volumes are increasing 4% to 8% annually, with new radiologists entering the field at a flat rate (roughly 1,050 a year, up only from 950 about 20 years ago). “As a result, every one of us is reading more and more studies every day," he said.

With workloads growing, radiologists need technology that removes friction rather than introducing new steps, clicks or complexity. "So, they really need things that just work with their existing workflows, especially as you start moving into more critical workflows," Dr. Chang said.

Meaningful innovation, he said, starts by understanding how radiologists move through their day, where inefficiencies exist and how technology can eliminate repetitive tasks without disrupting the reading process.

Building Around Radiologists, Not the Other Way Around

For Dr. Chang, firsthand clinical experience has been essential to developing technology that addresses real workflow challenges. "You gain a lot from understanding the workflow really, really deeply," he explained.

That understanding helped shape Rad AI's approach to generative AI. Rather than requiring radiologists to adopt an entirely new process, Rad AI initially focused on automating one of the most repetitive parts of reporting: generating impressions.

By creating impressions tailored to each radiologist's preferred language, the technology reduces manual dictation and editing while allowing radiologists to maintain their established workflows.

The same philosophy extends to Rad AI's broader platform: AI should support the entire reporting process, not operate as another disconnected tool that radiologists have to manage.

The Next Wave: Agentic AI and Faster Innovation

Beyond radiology workflows, Dr. Chang also discussed a broader trend reshaping technology development: the rise of agentic coding.

AI-powered coding tools are changing how software gets built, helping development teams move faster, experiment with new ideas and accelerate innovation. For companies like Rad AI, these advances create opportunities to develop and refine solutions more quickly while staying focused on the needs of clinicians.

Better AI Starts With Better Workflows

For healthcare organizations evaluating AI solutions, it is important to go beyond features and capabilities and ask more practical questions: Does the technology fit into existing workflows? Does it eliminate unnecessary steps? Does it make radiologists' work easier?

The takeaway from Dr. Chang's conversation is clear: The most valuable AI doesn't require radiologists to adapt to technology. It's designed to adapt to them.

Listen to the full Outcomes Rocket podcast episode.