Episode #164. Fighting Burnout And Missed Findings With A Unified AI View, with Lior Eshel Ansher, Founder and CEO of TestDynamics

Summary

Podcast Summary:

Risk Never Sleeps Podcast – Episode #164
Title: Fighting Burnout And Missed Findings With A Unified AI View
Guest: Lior Eshel Ansher, Founder & CEO, TestDynamics
Host: Ed Gaudet
Date: December 15, 2025

---

Episode Overview

In this episode, host Ed Gaudet sits down with Lior Eshel Ansher, founder and CEO of TestDynamics, at the AIMed25 conference in San Diego. The conversation centers on the critical issue of diagnostic errors in radiology, the mounting burnout among radiologists, and how a unified AI solution can help address these challenges by integrating diverse AI tools into a single actionable workflow. Lior shares his personal journey, sparked by his father's misdiagnosis with lung cancer, and discusses his vision for the future of AI in healthcare.

---

Key Discussion Points & Insights

Lior’s Background and Motivation

[01:18 – 03:03]

• Originally from Israel, currently living in Austin, Texas.
• Has a deep engineering and startup background, including mentoring and working with many early-stage companies.
• Inspired to enter medical technology after his father’s late-stage lung cancer misdiagnosis—his personal research revealed misdiagnosis rates of around 30%.
• Recognizes widespread issues in radiology: high error rates, increasing imaging volumes, radiologist shortages, and significant professional burnout.

"At one point during my research, I realized that 30% of the cases are misdiagnosed, which is a very high number. That's huge."
— Lior, [02:28]

The State of AI in Radiology

[03:03 – 05:53]

• Radiology is a leader in medical AI adoption with many specialized (narrow) AI models focused on detecting specific findings, such as brain bleeds or fractures.
• Radiologists are evaluated by speed and accuracy, but increasing demands and limited time lead to missed findings.
• The market is crowded: 1,200+ approved imaging AI solutions, each with a narrow focus.

"Each one of the AI tool[s is] being developed for something very specific... but when you as a radiologist read these cases, you need to look at the image in a holistic way."
— Lior, [03:20]

TestDynamics’ Unified Approach

[03:52 – 05:53]

• TestDynamics comes in post-read: after initial radiologist review, their solution aggregates multiple AI outputs into a single, unified dashboard.
• Their platform (Satori) allows clinicians to leverage numerous AI models without having to interpret disparate outputs.
• Focuses on making AI actionable and user-friendly for radiologists, helping them catch missed findings while reducing cognitive burden.

"We basically combine data from multiple sources and create one unified way to understand the outputs from different AI models we activated."
— Lior, [04:40]

Human-in-the-Loop: The Essential Partnership

[05:07 – 06:52]

• Lior highlights that AI is an augmentative tool, not a replacement. Human clinicians remain paramount as the ultimate arbiters of diagnosis.
• Envisions radiologists becoming “orchestrators,” akin to pilots who monitor an autopilot system, prepared to intervene in crises.

"I think the market is evolving into a human plus AI world..."
— Lior, [05:42]

"A good analogy is maybe if you are a pilot... today, it's mostly autopilot. But they sit there because they are there for crises and for things that they need to be aware of."
— Lior, [06:33]

The Future of Radiology & AI

[05:53 – 07:06]

• Predicts that in the next 10 years, radiology’s focus will shift: away from isolated image reading to the orchestration of multiple AI tools, synthesizing diverse data sources.
• The radiologist’s role becomes interpretation, oversight, and coordination, instead of direct “reading.”

Growth and Challenges as a Startup

[07:08 – 09:36]

• TestDynamics is in production and deployed across the U.S.
• Satori is their central product, focused on ease of integration for health systems.
• Big challenges include the slow pace of healthcare sales and adoption, the emergence of new decision-makers like Chief AI Officers and AI committees, and the need for efficient, unified integration of multiple AI tools.

"Slow sales cycles that everyone can complain about. The health system is moving very slow."
— Lior, [08:41]

"What we are offering... will provide you a piece of quiet time that you can understand or use one single framework to deploy AI."
— Lior, [08:53]

Personal Insights & Philosophy

[09:36 – 14:10]

• Lior reflects on Israel’s vibrant startup scene and the importance of supportive entrepreneurial communities.
• Shares his background as a musician and discusses the importance of creativity and innovation across disciplines.
• Expresses a deeply personal drive to address diagnostic errors after losing his father, Shmuel, to lung cancer—dedicates his work in part to his memory.

"I have a personal ambition to fix something here."
— Lior, [13:51]

Lighthearted Moments

[11:08 – 14:10]

• Lior admits that, if not an engineer, he might have been a singer or musician.
• Shares stories about his family, including missing his son's (Daniel's) birthday for the conference, and reminisces about his late father’s love of music and time as a DJ.
• Fun exchanges about favorite bands and albums—Beatles, Elton John, Muse, Foo Fighters, Dire Straits.

---

Notable Quotes

• On the challenge in radiology:
  "Radiologists... have less time spent on a patient case and they might miss something in study. So how AI can help here is... a good question."
  — Lior, [03:03]

• On AI’s evolving role:
  "AI would be a co-pilot at one point. A system they can use and leverage. But they will need to sit there and put the reports and orchestrate everything..."
  — Lior, [06:52]

• On entrepreneurship & community:
  "Every time people ask me why Israel? What's so special about Israel? We know how to communicate... and we know how to help each other over time."
  — Lior, [09:48]

---

Important Timestamps

• [01:18] Lior shares his personal journey and inspiration for TestDynamics.
• [03:03] Discussing today’s challenges in radiology and AI’s potential role.
• [03:52] Explanation of TestDynamics’ post-read, unified AI framework.
• [05:55] Predictions for radiology’s future – “the orchestrator” model.
• [07:08] How the company started – not an AI company, but an integrator.
• [08:41] Challenges selling into the healthcare sector.
• [09:36] Reflections on Israel’s start-up culture and what shaped his approach.
• [13:44] Heartfelt tribute to Lior’s late father, Shmuel, and dedication to preventing misdiagnosis.

---

Conference Impressions

[15:10 – 16:06]

• Lior is an active participant at AIMed25, moderating panels and showcasing practical AI use cases in radiology.
• Both appreciate Dr. Chang’s leadership and describe the conference as educational, community-focused, and refreshingly light on sales pitches.

---

TestDynamics & Satori Platform

[14:18 – 14:51]

• Company website: testdynamics.net
• Platform name: Satori, meaning “enlightenment” in Japanese—selected to reflect the aim of bringing clarity to radiologists’ work.

---

Final Thoughts

Lior Eshel Ansher demonstrates a mission-driven approach to healthcare innovation, marrying personal experience with a technical vision for AI as a supportive, human-centered tool in radiology. The conversation covers the real-world pain points of clinicians, the overlooked barriers to AI integration, and offers a hopeful, practical vision for the future of medical imaging.

---

Quick Links

• TestDynamics Website
• To learn more, visit censinet.com

---

This summary covers the essential content, skipping introductions and advertisements for clarity and depth.

Transcript

[A]

Foreign. Welcome to Risk Never Sleeps, where we meet and get to know the people delivering patient care and protecting patient safety. I'm your host, Ed Gaudet. All right. Ed Gaudet from aimed25 conference in San Diego sponsored by Outcomes Rocket and Senseinet. Risk Never Sleeps. And I am here with my good friend. My God, what is your name? We were running to get this one. Lior. Lior. All right. Eschel from Test Dynamics.

[B]

Yes, you're right.

[A]

All right, now you're from Amsterdam.

[B]

No, I'm originally from Israel.

[A]

Israel. Oh, okay. Yeah, right. I completely messed up the whole thing. So this is going to be really an interesting pot. That's a good sign, actually, when you mess it up in the beginning. Yeah, he's not buying it.

[B]

I'm also hosting a podcast, so. I know.

[A]

Are you.

[B]

I know exactly what it's like. Yeah.

[A]

And we have. I think we're doing like 30.

[B]

Oh yeah.

[A]

In the next three days.

[B]

Yeah.

[A]

Maybe even more. I don't know.

[B]

That's it.

[A]

But we're going to leave you with a nice gift, so for all your trouble.

[B]

Okay. Thank you so much.

[A]

All right, so tell us a little bit about yourself, your background, your company.

[B]

Yeah. So I'm Liorashel, I'm originally from Israel, currently live in Austin, Texas. Married, two kids. My background is engineering. I am. I'm an engineer. I work for industry for a long time, either corporate startups. I moved to launch my own company before this one which was help companies build products and launch it to the market in A to Z product development. Oh wow. We work with many startups, early stage mentor, many startups, been part of many startups evolution. And then at one point in my life my dad was misdiagnosed with lung cancer and it was a shocking event in my life and I tried to understand how this type of mistakes happens and misdiagnosis. And I did my own research on radiology and tried to understand how they read, what are the challenges they are facing day to day and how technology can help them overcome and maybe reduce the errors. At one point during my research, I realized that 30% of the cases are misdiagnosed, which is a very high number. That's huge. So my dad is not the only one. And these mistakes happen every day in hospital setup and specifically radiologists. They are the front leader with technology. They use technology all the time. They always improved the technology. They moved to be digital 20 years ago and they continue evolving in the space and now they implementing AI, which is, they are, I think it's one of the most mature market for AI. But the problem is that the challenges they are facing day to day are big challenges. First of all, there is a huge burnout and shortage of radiologists today. On the other hand, they have, they're facing a lot of increasing volume. Right. A lot of images they need to read in that place.

[A]

Yeah.

[B]

And they've been measured by time. Right. How much time it takes you to read and analyze the case and the rvu, the threshold just go up and up every year. So they have less time spent on a patient case and they might miss something in study. Yeah. So how AI can help here is. It's a good question. Because AI, there are many narrow solutions in this market. Each one of the AI tool being developed for something very specific. Right. To find the bleed in the brain or to find a fracture in the X ray. But when you as a radiologist read these cases, you need to look at the image in a holistic way. You need to may activate more than one AI engines on the specific case to be able to find different abnormalities.

[A]

So are you coming in post read or pre read?

[B]

So we come in post read. So Test Dynamics basically started as an intermediate company and approach. We like to know the market. We like to basically think about the end users and how they consume the data in an actionable way. How they can integrate AI and have AI make the holistic view on the patient study and help them understand what's going on with the patient. Which means you may need to activate more than one AI engines on a specific study. And you don't want to, you don't want to read different outputs every time you integrate a new AI. So what we do, we congramulate the data. We basically combine data from multiple sources and create one unified way to understand the outputs from different AI models we activated.

[A]

Interesting. And are you in production? Do you have customers?

[B]

Yes, we are in production. We have customers.

[A]

Us.

[B]

Yeah, we focus on the US market. Okay.

[A]

We're all over the US or right.

[B]

Now it's all over.

[A]

Nice. That's great. Congratulations.

[B]

Yes.

[A]

And how long has the company been around?

[B]

So we are, we established the company about four years ago.

[A]

Okay.

[B]

We spent a lot of time on building this intelligent framework.

[A]

Yeah.

[B]

How we deploy AI, how we understand the outputs, how we understand the users, how do we combine the two and make it useful. And again, people talk about AI going to replace clinicians and radiologists for many years. I think the human in the loop is a very big aspect of it and we need them to be advocate to AI and we need them to change their role to be able to accept the AI and work alongside it. But eventually they will override AI and they will basically be the clinicians that we put a report in front of the referring physicians and the patients. So I think the market is evolving into a human plus AI world which is basically will allow them to find incidental findings that they might miss or. Yeah.

[A]

What's this going to look like in 10 years?

[B]

In 10 years I think they, they will move from the way they're reading today. Today they are reading and they need to be very concentrated when they read the study. I think at one point AI will be able to read or at least give them a preliminary report on the study, what's going on with the patient by combining multiple sources and create this to one unified way, unified interface. But eventually I think they will move to more being orchestrators than readers.

[A]

Okay, so the job, the reader job you think goes away and becomes something else.

[B]

Yeah, it becomes more like an orchestrator. Orchestrator. I think a good analogy is maybe if you are a pilot, right. I don't know, 20, 30 years ago pilot really land and take off the plane. Today it's mostly autopilot.

[A]

Yeah, Right.

[B]

But they sit there because they are there for crisis and for things that they need to be aware of.

[A]

Interesting.

[B]

So I think radiologists would be AI would be a co pilot at one point. A system they can use and leverage. But they will need to sit there and put the reports and orchestrate everything and be and control the AI use over time.

[A]

Interesting. And are you the founder?

[B]

Yes, I'm the founder.

[A]

Come up with the idea.

[B]

So during my research back then I talked to many audiologists, many physicians and I tried to understand their challenges. And then I real that there is a fascinating market on the other side, which is AI companies that are just forming every day. And I realized that this market will grow exponentially over time. And you will see more and more companies getting clearance by FDA. Today over 1200 AI applications only in imaging, which is only imaging. Yeah, only imaging, which is a huge number. And if they just keep regulated them, maybe tomorrow it will be 1400, I don't know. Right. Yeah.

[A]

Or 5000.

[B]

Yeah, it just keep growing. So I, I, you know, we, I analyzed the market and I realized that this wouldn't go continue growing. Right. But the problem is they how do you make it all useful for the physicians and actionable? And that's the challenge. We try to solve it. Not we're not an AI company. We're trying to know AI very well and I understand, but we try to be the shovels.

[A]

Yeah. Now this is your second third company you started?

[B]

It's my second company, yeah.

[A]

What did you learn from your first that you brought over?

[B]

The first company was not really a startup. It was. Yes, forming from scratch. I was an engineering house support company. Okay. We worked with many different companies. We worked many startups. I was actually fractional CTO from many startups and PPR and D and I work in many different startup environments. So I didn't experience the A to Z how to scale a company. And that's what I'm using here as well.

[A]

Interesting. What are some of the biggest challenges you're facing right now?

[B]

Slow sales cycles that everyone can complain about. The health system is moving very slow. There are new positions in health systems today like the chief AI officers or the AI committees which formalizing over time.

[A]

Right.

[B]

So that's a challenge. The other challenge might be discussions with IT leaders and the CIOs and the CMIOS. I think the eventually when you have something like what we are offering, it will provide you a piece of quiet time that you can understand or use one single framework to deploy AI. So even on the business model it's something very interesting and it's just a one line in your budget. So you don't need to deal with all these different AI vendors and the 1200 applications I told you about. You just need to understand what's your clinical use cases you want to activate and what's the value you want to get from AI. And then we will be your partner to help you understand this market and deploy it.

[A]

If you can go back in time and tell your 20 year old self something, what would you tell him?

[B]

I think I was always like strive to admission I had this entrepreneurship spirit. I think being in Israel it's one of the centers, one of the centers for startups. I think we have like 20 startups lunch every week which is like incredible. It's credible. So I think being in this environment, I think being Israeli is also helping us creating communities and get support from people. And every time people ask me why Israel? What's so special about Israel? We know how to communicate with each other and we know how to help each other over time. So you always have someone to ask something if you have a challenge. So that's something about us being Israelis. But I think even when I was an employee for corporate companies, I always like to develop and innovate and launch some something new. Even when I work for Motorola and it was like really hard, you know, to move this ship, but I managed to build an innovation team and I managed to do something as a young engineer. So I do see this. These capabilities now forming with this company. And I think we are creating a new category in medical imaging, which is the amalgamation of the data.

[A]

The orchestrator.

[B]

The orchestrator. Not just the orchestrator, but also the conglomeration of the data, extend outputs and create something that is more useful eventually for radiologists.

[A]

Interesting. If you could go back in your life and change one decision you made.

[B]

What would it be? Oh, wow. That's a good question.

[A]

Yeah, I just made it up. I don't think I've ever asked that question.

[B]

I don't know. In the past. I was. Before I went to engineering school, I was also a singer. I used to play piano. So I had this Billy Joel. I questioned my.

[A]

Or Ellen John.

[B]

No, Ellen John.

[A]

Ellen John.

[B]

Yeah. But I questioned my. I questioned myself if I was. If I need to be singer or an artist or an engineer. So that was like a.

[A]

But you sang in public right now?

[B]

No, no. But you did. Yeah, I did. Yeah.

[A]

You do karaoke?

[B]

No.

[A]

Ever? No, never.

[B]

I did some with friends.

[A]

By the way, you know the song, your song by Elton John?

[B]

No, I don't think so.

[A]

It's a little bit funny.

[B]

Oh, yeah, of course.

[A]

You know that song.

[B]

Yeah, of course.

[A]

Oh, you're not going to sing?

[B]

Oh, no, But. But I like to do. I like to play music and I am now learning guitar and other stuff. But. Yeah, but the music is my.

[A]

You'd be doing music if you weren't doing this.

[B]

Exactly. Yeah. So that's my hobby. Oh, all right.

[A]

I love that desert island. You could bring five records with you. What would you bring?

[B]

Five records? Yeah, five albums, I think I like the Beatles a lot and I like. You mentioned. Until John and Beatles.

[A]

Elton John.

[B]

Yeah, yeah. And I like also some rock music. The Muse.

[A]

Muse.

[B]

Yeah.

[A]

Nice.

[B]

Yeah. Foo Fighters. Oh, Foo Fighters. Ye.

[A]

One more.

[B]

One more. Yeah, make it a good one.

[A]

No pressure.

[B]

I don't know one. No, no jazz.

[A]

No classical.

[B]

Maybe in Dire Straight. Yeah, that was.

[A]

Dire Straits is the best one.

[B]

That was my. My dad's like, favor band.

[A]

Yeah.

[B]

You went to two shows in Israel and then he came to Israel.

[A]

Really?

[B]

Yeah.

[A]

I love Dire Straits, so.

[B]

Yeah.

[A]

Oh, yeah.

[B]

Yeah.

[A]

That was your dad's favorite brand.

[B]

Yeah, my dad's my favorite.

[A]

Great.

[B]

My dad was a DJ when he was young. Really? Maybe that's the music. That's where the music. Oh, yeah, yeah.

[A]

And you said you have two kids?

[B]

Two kids, yeah.

[A]

What do you have?

[B]

A girl and a boy. Oh, old. Nine and seven.

[A]

Great ages. You're in the middle of it, right? Oh, don't you want to just, like, stop time?

[B]

Yeah, I do. And actually, my son's birthday is today, and I'm missing it.

[A]

What's his name?

[B]

Daniel.

[A]

Like the song? Like the yachts.

[B]

Like the song, yeah.

[A]

Did you name him after the Elder?

[B]

No, but I. I did. He knows this song.

[A]

Happy birthday, Daniel. If you're listening to the podcast, your dad is an amazing guy. If he's not your hero, he should be your hero, right?

[B]

Yeah.

[A]

Aren't all son. Aren't the fathers of all sons their hero? I think. Right?

[B]

I agree. Yeah. Yeah, yeah.

[A]

Is your dad still with us?

[B]

No, my dad passed away two years ago.

[A]

I'm so sorry.

[B]

Yeah, he was. He was misdiagnosed and diagnosed very late.

[A]

You mentioned that.

[B]

With lung cancer.

[A]

My mom died from lung cancer.

[B]

Yeah. Yeah. So that's an awful. Yeah. But I. I have a personal ambition to fix something here.

[A]

And you'll be successful because of that.

[B]

Yeah, I hope so.

[A]

Yeah. What was your dad's name?

[B]

Shmuel.

[A]

Shmuel.

[B]

Yeah.

[A]

Oh, well, we'll dedicate this podcast to Shmuel.

[B]

How about that? Thank you so much.

[A]

And happy birthday, Daniel. And it's been really great to meet you.

[B]

Yeah. Great to meet you.

[A]

How can people learn about your company?

[B]

First of all, you can go to the website, test dynamics.net. okay. Our platform name is Satori and. Satori. Oh, yeah, from the Japanese. I know. I probably know the meaning, but Satori means sun and enlightenment or a moment of clarity. And this is what we try to bring to radiologists.

[A]

Writers wrote about Satori.

[B]

Yes.

[A]

Jack Kerouac.

[B]

And you're right. Yeah.

[A]

Did you ever read that book Satori?

[B]

No, I don't read that.

[A]

That's pretty interesting. No, I love the name of your product. And yeah, I think your product will become your brand because Test Dynamics, not so good. But Satori.

[B]

Yeah, Satori is great.

[A]

That's an interesting brand. I love that. Yeah. All right, well, check out Test Dynamics. Check out the product. Satori, any last comments for guests? I mean, you're an interesting. I could spend all day with you.

[B]

No, I think I really enjoy here being an AI man.

[A]

It's been a great show, hasn't it?

[B]

It's my second time here.

[A]

What was last year like?

[B]

Last year was good. I mean, we also as big as this year. It was. Yeah. The same size.

[A]

Yeah.

[B]

This year. So last year I did one session with Radiologist. We actually showcase a few use cases in the show. Nice. And this year I am moderating two panels and another one, I'm studying another one. So. Yeah. So it's great that they are bringing the AI and medical imaging sector to this conference. And I really like Dr. Chang, which is an amazing guy.

[A]

Dr. Chang is an amazing, amazing person. Amazing man.

[B]

Right.

[A]

All he does for the community and the industry.

[B]

Amazing. Yeah.

[A]

This is the. One of the. I think it's probably the best AI conference.

[B]

Yeah.

[A]

Great content. Right. Not a lot of glitz. Right. You have some vendors here, which is good.

[B]

Yeah.

[A]

But you don't have the. It doesn't feel like a sales pitch. Right. It feels like you're learning, always learning every minute of this conference. Yes.

[B]

And people are very nice and also very nice. Dr. Chang is a tool giver. Is giving a lot of the community.

[A]

Absolutely. He's so generous with everything. All right, well, we're going to wrap it up. Lior, thank you very much for your time. You've been a great guest and I wish you the best of luck.

[B]

Thank you so much. Thank you.

[A]

Thanks for listening to Risk Never Sleeps. For the show, notes, resources and more information and how to transform the protection of patient safety. Visit [e-mail] that's sensitivity. C E N S I N E T Com. I'm your host, Ed Gaudette. And until next time, stay vigilant because Risk never Sleeps.