Fresh off RSNA, one thing was clear: radiology has officially moved past the era of “AI as a demo” and into “AI as infrastructure.”
For years, the exhibit hall has been filled with impressive algorithms and detection tools. This year felt different. The conversation shifted from shiny widgets to real workflows—how AI fits into complex clinical environments, integrates with legacy systems, and actually supports patient care end to end. Technology selection is no longer about liking a tool. It is about vision, interoperability, budget reality, and operational readiness. Radiology leaders are now asking tougher questions:
- Can outcomes be measured?
- How will this integrate into what already exists?
- What happens after a finding is detected?
At the same time, buyers are getting sharper. Hype is giving way to proof. Leaders want clinical results, workflow transparency, and client references—not marketing promises. If a vendor hesitates to show real-world outcomes, that is a red flag.
Another major shift is the rise of multimodal and agentic workflows. Radiologists want imaging, labs, clinical context, and EMR data connected—so decisions can happen faster and with more confidence. But none of that matters without a follow-up plan.
One of the most sobering truths from RSNA: detection without care orchestration increases risk. A worklist is not follow-up. Turning on algorithms without a coordinated pathway for patients creates liability, not value.
The future is not closed platforms or disconnected tools. The future is interoperable, workflow-driven systems that ensure patients actually receive the care their images identify.
At RSNA, one message came through loud and clear: interoperability is no longer a “nice to have.” It is now the competitive advantage.
Episode chapter guide:
01:50 Insights from RSNA: Technology and Quality Improvement
04:41 Navigating Technology Selection in Radiology
08:39 The Evolution of AI in Radiology
12:23 Interoperability and Workflow Integration
17:59 Demand for Proof: Moving Beyond Hype
20:33 Care Orchestration and Follow-Up Challenges
Full Transcript
AI-generated transcript. Accuracy may vary; please excuse any transcription errors.
Angela Adams, RN:
Welcome to After the Chaos with Angela Adams. I’m the CEO at Inflo Health, and today I’m joined by Kandice Garcia, founder of Tungsten QI Partners.
Kandice Garcia, RN:
I’m Kandice Garcia, founder of Tungsten QI Partners.
Angela:
We are fresh off our return from RSNA and wanted to give our takeaways and the fresh scoop for those who weren’t able to fight the blizzard and come to Chicago. Kandice, it sounds like you got the RSNA bug.
Kandice:
I did. I’ve been sick for the last four days. Luckily I had a quiet weekend to recover, but I don’t think I’m the only one.
Angela:
You’re the third person I talked to today with the RSNA bug. It’s a rite of passage.
Kandice:
It is. This was my ninth year attending, and it was one of the busiest and most exciting years yet—not just in content, but in evaluation and relationships. I have so many thoughts swirling.
Angela:
Tell me what made it exciting and give me your view on hype versus reality.
Kandice:
One of our biggest focuses at RSNA was technology selection. We did concierge selection with radiology leaders—walking the floor with them, helping them plan their time, ask the right questions, organize the information, and understand what they were seeing.
What became very clear is that decisions today are far more complex than liking a tool or an algorithm. Leaders have to think about vision, platforms versus PACS, legacy systems, budgets, and whether they can buy everything or have to prioritize.
We also know that if you can’t measure a process, you can’t improve it. Leaders had a lot of questions about data, monitoring, and proof. Often, the answers were unclear. The tech is improving, but the data is still behind where leaders need it to be.
Everyone claims they improve efficiency and the bottom line. It’s hard to discern marketing from reality. For leaders entering this AI world, it’s very hard to know what to trust.
Angela:
Some demos are vaporware—future-state products that aren’t live yet. Buyers don’t always realize they’re essentially beta users.
Kandice:
Exactly. We pushed vendors to walk through real workflows. We asked how implementation actually works in organizations that are messy and variable. Often, there was no clear implementation plan.
One thing that helped was talking with other radiology leaders outside the showroom. We gathered real feedback from people actually using these products. That helped us surface limitations and ask better questions.
Our role now is helping leaders organize everything they learned and make thoughtful decisions over the next few months.
Angela:
From our side, Inflo had the busiest booth we’ve ever had. We were booked solid, even with the weather. The AI Showcase has exploded compared to four years ago.
What stood out most is that AI has moved from widgets to infrastructure. Companies are now focused on how AI fits into workflows and integrates with legacy systems.
Radiologists want multimodal, agentic workflows. They want to interact with systems more naturally and pull together imaging, labs, and clinical context.
Kandice:
That evolution makes decisions harder. Leaders aren’t just choosing algorithms anymore—they’re choosing platforms and even considering replacing entire systems.
But flexibility is key. Leaders want workflows, not algorithm bundles. They want soup-to-nuts lung cancer screening, CAC, liver disease—end-to-end.
Angela:
Interoperability is now the competitive advantage. Closed systems will fall behind. AI should fit into existing workflows, not force hospitals to rebuild everything.
Hospitals are also demanding proof, not hype. Leaders want access to real clients, real workflows, and real results.
Kandice:
And none of this matters without follow-up. A worklist is not follow-up. Turning on detection without care orchestration creates risk.
We’re collecting massive amounts of information with no clear path for action. We need centralized platforms and clear handoffs.
Angela:
We see this constantly. Health systems turn on detection and then realize they don’t have the infrastructure to manage what they’ve uncovered. Follow-up has to come first.
Kandice:
Technology without organizational design will fail. Workflows, responsibilities, and handoffs all have to be clearly defined.
Angela (closing):
If you need help with technology selection, call Kandice. If you need help with follow-up, call us. RSNA, we love you—we just don’t love Chicago in December.
