AI ROI: What Early Adopters Are Measuring and Why It Matters

Summary

Podcast Summary: Becker’s Healthcare Podcast

Episode: AI ROI: What Early Adopters Are Measuring and Why It Matters
Host: Lucas Voss, Becker's Healthcare
Guest: Alex LeBrun, Co-founder and CEO of Nabla
Date: December 17, 2025

---

Overview

This episode explores how early adopters in healthcare are defining, measuring, and maximizing ROI (Return on Investment) for AI implementations. Lucas Voss interviews Alex LeBrun, the CEO of Nabla, who shares insights into the evolving methodologies, metrics, and leadership approaches as organizations move from pilot projects to system-wide AI adoption. The discussion focuses heavily on revenue cycle management (RCM), scalability, and the transition toward more tangible, financially measurable ROI in clinical workflows.

---

Key Discussion Points and Insights

Introduction and Alex LeBrun's Background

(00:29-01:19)

• Alex LeBrun: Co-founder and CEO of Nabla, focused on building AI assistants for clinicians to reduce administrative burden and paperwork.
• Nabla’s reach: Serving ~85,000 clinicians and processing millions of encounters monthly.

The Shift from Soft to Hard ROI in AI

(01:49-02:59)

• Phase One: Initial AI deployments in healthcare had "soft ROI": improving clinician happiness, reducing burnout, and “pajama time,” but without clear financial impact.

• Phase Two: Now entering a phase where "very strong financial ROI" is expected and measurable.

  “If you are like cynical, you could say, hey, the CFO didn’t see the impact on their numbers...with more maturity in the field...we need to have very, very strong financial ROI when we deploy AI.” – Alex LeBrun [01:59]

• Measuring ROI is now precise and rigorous, a change from as recently as six months prior.

What Early Adopters Measure for AI ROI

(03:31-04:27)

• Main focus: Impact of AI on claims and revenue cycle management (RCM).

  • Many clinicians under-code due to audit fears or uncertainty, which leads to lost revenue.
  • AI assistants help clinicians code more accurately and fully, leading to immediate, measurable revenue increases.

• Trend: Most ambient AI companies are quickly integrating with RCM because ROI here is “immediate, easy to measure.”

  “The most obvious [impact] is…on revenues...you can really measure the impact of having this AI assistant helping the clinician to code as they are supposed to code, which results in revenues directly.” – Alex LeBrun [03:41]

How Health Systems Are Measuring ROI & Scaling Initiatives

(04:27-05:54)

• Measurement practices are very new; systematic ROI measurement started only recently.
• Health systems are converging on good KPIs and metrics, but LeBrun urges caution—it is early and results are still forthcoming.
  • Advocates for humility and ongoing evaluation over hype.

  “A few months ago no one was measuring this kind of ROI and we just started to set up the right metrics, KPIs to see how AI impacts revenues…We should meet again in six months and see what was actually measured.” – Alex LeBrun [05:17]

Leadership, Change Management, and the Path to True ROI

(05:54-08:00)

• True ROI requires integration—AI must not live in silos but connect across workflows.
  • Hurdles: Point solutions are still common, but move toward “end-to-end solutions” is accelerating.
• LeBrun envisions AI managing processes from pre-charting to reimbursed claim, making ROI “easy” and “obvious.”
• Progress toward agentic, automated, connected systems will make impactful metrics more transparent and achievable.

  “Once you have an end-to-end solution…measuring very strong ROI is easy because you can…prove that it actually had an impact two months later on revenue collection.” – Alex LeBrun [07:02] “We are excited because this kind of ambition to have an end-to-end process workflow managed by AI was impossible to even imagine a year ago.” – Alex LeBrun [07:23]

Final Thoughts and Looking Ahead

(08:29-09:22)

• AI in healthcare is at the very beginning—comparable to the early days of the internet.
  • Success will demand customization and modeling for each client—AI is not an "off-the-shelf" product.

  “It’s just the beginning…maybe like if we were talking internet in 97. So it’s not perfect, many things are still in progress…It’s 1% journey done, 99% to go.” – Alex LeBrun [08:29]

• The future conversation isn’t “should we use AI,” but rather “how can we rethink workflows with AI’s new capabilities?”

---

Notable Quotes & Memorable Moments

• On AI’s Impact on Clinicians’ Work:

  “Our goal has always been to build an AI assistant for clinicians, particularly to help them get rid of the part of their jobs they don’t like...” – Alex LeBrun [00:41]

• On Phase Shift in Measuring ROI:

  “With more maturity in the field...now we are measuring very, very precise ROI. And it's new, actually. It was not like that just six months ago.” – Alex LeBrun [02:37]

• On Moving to End-to-End Solutions:

  “What’s exciting now is...we’ll see more and more like end-to-end solutions built on AI. And once you have...everything...powered by AI, then measuring very strong ROI is easy.” – Alex LeBrun [06:44]

• On Perspective for the Industry:

  “People should be aware that it’s just the beginning...the discussion we have shouldn't be ‘should we use AI or not’...it's ‘how can we rethink this workflow given the new tools we have?’” – Alex LeBrun [08:37]

---

Important Segment Timestamps

• Introduction & Guest Background (00:00–01:19)
• Emergence of Hard AI ROI in Healthcare (01:49–02:59)
• Top ROI Metrics & Revenue Cycle Management (03:31–04:27)
• Real-World Measurement Practices (04:27–05:54)
• Leadership, Change Management, and the Future of AI ROI (05:54–08:00)
• Final Thoughts: Looking Ahead in AI for Healthcare (08:29–09:22)

---

Conclusion

This episode provides a candid and forward-looking examination of how healthcare is evolving in its approach to AI ROI. Moving from soft benefits (like clinician satisfaction) to quantifiable revenue impact, health systems and vendors are beginning to rigorously define and measure ROI, particularly in revenue cycle management. Alex LeBrun emphasizes humility and continuous adaptation as this journey progresses, urging leaders to envision not just if, but how, AI can redesign their workflows for greater organizational and patient benefit.

Transcript

[A]

Hi, everyone, this is Lucas Voss with Becker's Healthcare. Thanks so much for tuning in to the Becker's Healthcare podcast series. Fantastic to have you all. And an exciting topic today. Very top of mind for a lot of folks. AI roi, what early adopters are measuring and why it matters. And I'm even more excited to be joined by Alex Lebrun. He's the co founder and CEO of nabla. Alex, it's so great to have you. Thanks for being here.

[B]

Hi, Lucas. And thanks very much for having me today.

[A]

Yeah, absolutely. I do want to start off with introductions for those that might not know you yet. Could you just share a little bit about yourself and your work in healthcare?

[B]

Yeah. I'm Alex LeBrun, co founder and CEO of Nabla. I started Nabla seven years ago. Our goal has always been to build an AI assistant for clinicians, particularly to help them to get rid of the part of their jobs they don't like and where they are wasting a lot of time, administrative stuff, paperwork. So, yeah, we've been working on that at NABLA for many years now, reaching today about 85,000 clinicians. Millions of encounters every month. And we have a very ambitious plan to do more and more for our clinicians as we move forward.

[A]

I recently had the pleasure to listen to you at our Becker CEO and CFO roundtable where you're part of a panel and the discussion really centered around how early AI adopters like UNC Health Penn Medicine are. Are defining and measuring ROI right now. When you look back at that conversation, what surprised you most as you were listening to your peers and the conversation itself and, and on the other hand too, what validated what you were seeing and what's happening right now?

[B]

Yeah, so. So what's surprising and amazing is that we really are starting a new phase. Until recently, we were, I think, in phase one of AI deployment in healthcare, where we didn't have hard roi. It was a like soft roi. So making physicians happier, less burnout, less pajama time. People are really happy to use AI. It has an impact. But if you are like cynical, you could say, hey, the CFO didn't see the impact on their numbers. And it was phase one. And I think with more maturity in the field and also with bigger projects, more investments, now we are starting phase two where we need to have very, very strong financial ROI when we deploy AI. And it was really obvious when you listen to the panel, to my partners at the panel, that now we are measuring very, very precise roi. And it's new, actually, it Was not like that just six months ago.

[A]

Yeah, we talk so much about the transition from pilot to scalability, right. Organizations now, scaling these initiatives across organizations and obviously measuring ROI around them. And there's so much of a range for ROI priorities right now. Mission alignment, administrative burden is certainly one that's top of for a lot of people. Patient trust, lots of different priorities. From your perspective, what are some of the measures 1, 2 or 3 that health systems should be prioritizing right now.

[B]

And why so in my field, you know, in our field, the most impact is, the most obvious is the impact the tool has on claims, you know, on, on revenues, so on. In terms of renew cycle management, you really see how many clinicians are actually down coding because they are afraid of audits. They're not sure about something, so they tend to miss a lot of things they should do. And you can really measure the impact of having this AI assistant helping the clinician to code as they are supposed to code, which results in revenues directly to revenues translated to revenues. So this is, and this is why in our field, most ambient AI companies are moving fast into coding and rcm, because this is a very, very strong, immediate, easy to measure kind of roi.

[A]

When you think, and I agree with the RCM piece, I think that's a lot of conversation that's happening right now. We saw it at the conference too, at the roundtables. A lot of it was RCM focused. When you think about the partners that were, some of your partners that were in Chicago there at the roundtable, can you share some of the examples of health systems that got measurement right of thinking about, okay, this is how we measure ROI and how did their approach influence also leadership buy in and how these initiatives then scale across organizations?

[B]

So that's a good question. I cannot speak for them and I don't remember their exact results and it's very new also. I think we should also not exaggerate too much that, okay, we've been doing that for two years and we have these numbers. I think I can say that a few months ago no one was measuring this kind of roi and we just started to set up the right metrics, KPIs to see how AI impacts revenues, for instance. So I think the methodologies we've converged to are good and we should meet again in six months and see the actual what was actually measured. I think it's still early. We should, we should be humble. It's still, still very early, but it's definitely the trend to go. There is, I think I Think we say to stay. Yeah.

[A]

Turning this around though to, to the leadership piece because I think again, change management when we talk about technology and AI is so key because people have to understand that they have to know the benefits that it brings. Right. And as we're looking ahead, how would you like to see healthcare leaders conversations about AI ROI evolve in the next couple of years and what are some of those opportunities that you are excited about?

[B]

Yeah, I think for an AI tool, if you live in a silo, in a sandbox, disconnected from the rest of the workflows, it's very hard to have a strong roi. That's why what we do is so difficult. What's exciting now is that as we have more coverage with AI, so there are still point solutions, but I think things are in the process to maybe consolidate and we'll see more and more like end to end solutions built on AI. And once you have an end to end solution going from the pre charting all the way to reimbursed claim and everything is powered by AI, then measuring very strong ROI is easy because you can follow, okay, I did something at the pre encounter stage, very, very early stage. But because I see all the process, I can prove that it actually had an impact two months later on revenue collection. You see what I mean? So we are excited because this kind of ambition to have an end to end process workflow managed by AI was impossible to even imagine a year ago. And with all the progress we have with agentic approach where we are not afraid to do more automation and to connect different tools together, then we will quickly reach this kind of end to end deployments. And for these ROI will be so obvious that metrics will be very, very obvious. So I'm personally quite excited by this perspective.

[A]

Well, and what you just mentioned too is so important for scalability, especially as we think about larger organizations, multiple locations, all of these different components. It's such a key to be able to scale these initiatives for organizations, which is awesome. Alex, it's so great to have you on. I do want to leave you with the floor for the final part of our conversation. Anything else that we didn't touch on? Any final thoughts that, that you'd like to share with the audience?

[B]

No, Generally I think people should be aware that it's just the beginning. We should be very excited because this is the beginning. Like if we were talking Internet in 97 maybe. So it's not perfect. Many things are still in progress and we love working with partners and customers who really realize that it's not an off the shelf thing that they have to take or not. It's something we have a very strong impact on modeling, on developing. And I really like to say it's 1% journey done as the journey is done. And the discussion we have shouldn't be should we use AI or not? For this is how should we can we rethink this workflow given the new tools we have and the new powers we have with AI? So this is this mindset I think great people have in our industries.

[A]

Absolutely. Alex, it's so great to have you. Thanks for your time and insights today.

[B]

Thank you very much, Lucas.

[A]

Absolutely. And we also want to thank our podcast sponsor, nabla. You can tune into more podcasts from Becker's Healthcare by visiting our podcast page at beckershospitalreview.