Join Megan Antonelli and Dr. Tom Kelly, co-founder of Heidi Health, as he brings a unique perspective to healthcare AI. He's a vascular surgery trainee turned founder who experienced firsthand the crushing administrative burden that keeps clinicians from truly connecting with their patients. With Heidi Health now serving over 500,000 clinicians and processing more than 2 million patient visits per week, Tom shares the fascinating journey from frustrated doctor to AI innovator. We dive deep into the reality versus the hype of ambient AI adoption, why clinician-first design trumps seamless integration, and the critical difference between tools that doctors tolerate and tools they actually love. From the nuances of personalization at the sentence level to envisioning an AI care partner that handles follow-ups and coordinates care, this conversation reveals what it truly takes to build healthcare technology that makes medicine more human—not less.
Dr. Thomas Kelly, Co-founder & CEO, Heidi Health
Megan Antonelli, Chief Executive Officer, HealthIMPACT Live
0:01 - Intro: Welcome to Digital Health Talks. Each week we meet with healthcare leaders making an immeasurable difference in equity, access, and quality. Hear about what tech is worth investing in and what isn't as we focus on the innovations that deliver. Join Megan Antonelli, Janae Sharp, and Shahid Shah for a weekly no BS deep dive on what's really making an impact in healthcare.
0:29 - MEGAN ANTONELLI: Hi everyone. Welcome to Health Impact and Digital Health Talks. Today we are here with Tom Kelly, the co-founder of Heidi Health. Heidi Health is a company redefining what's possible, an ambient AI for healthcare. Tom brings a rare perspective combining deep healthcare technology expertise with a clinical first mindset. With a mission to make healthcare more human through smarter, more intuitive tools, he and his team are proving that AI can seamlessly integrate into care delivery without disrupting clinical flow. Today we'll dive into the nuances of ambient AI adoption, the gap between technological promise and clinical reality, and what it takes to achieve the holy grail of healthcare technology, tools that clinicians actually want to use every single day. Well, wouldn't that be nice? Hi, Tom, how are you?
1:20 - TOM KELLY: Hi Megan, thanks for having me on the show and yeah, excited to talk to you.
1:24 - MEGAN ANTONELLI: Yeah, so excited to be speaking with you all the way from Melbourne.
TOM KELLY: Yeah, yeah, exactly, luckily we could find a good time for both of us, so I know it works, it works.
1:35 - MEGAN ANTONELLI: It's, it's your tomorrow.
1:37 - TOM KELLY: I'm still stuck here in today, but in the future, you know.
1:39 - MEGAN ANTONELLI: Yeah, yeah.
1:44 - MEGAN ANTONELLI: So, you know, we always like to start kind of hear about, you know, when we talk to founders on the program, kind of, what was your, you know, what, where were you and how did you come to found Heidi Health?
1:57 - TOM KELLY: Yeah, yeah, so, so yeah, I'm, I'm Tom, I'm a doctor by background, I, I loved working as a doctor, so I was training as a vascular surgery trainee, and, you know, got sort of 3-4 years out of medical school. I think, so before medicine I was one of those STEM kids that read all of the sci-fi books, you know, whether it was Dune or Asimov, and I was always like thinking of technology and before medicine I studied some computer science, so I was always thinking about machine learning, AI, building software tools on the side, and I think at the same time as that interest, I was seeing that as a doctor, I was having these experiences where I was just completely constrained in my ability to be a great doctor for my patient. There was so much paperwork, sometimes I would have to see maybe 100 patients between a few doctors in an afternoon, so it's like, and these patients had waited sometimes a year to see us, and it's a really important decision, you know, whether they have a surgery or not. And I was spending maybe 10 of the 15 minutes just trying to figure out like why are they here, have they got their documents, what do they need to book the surgery, the referral letter I have to write, what guideline do I have to look up, I have to remember to call them in a week. It's just like all this noise.
3:18 - MEGAN ANTONELLI: I mean you have the same, the same problems in Australia that we have here, huh?
3:24 - TOM KELLY: Yeah. So it was basically like I went, you know, I heard the calling to medicine because I loved the doctors in my life, like they just seemed like the most compassionate, the smartest people that I'd ever met, and they were kind, so they were the kind of person I wanted to be. But then when you're in practice, you try to do your best, but you just feel like you could always be being a better doctor every day and you're not giving your patients eye contact and giving them your best. So, yeah, I think when I was seeing some of, especially when Transformers came out in 2019, I was starting to experiment with different tools and was starting to realize, wow, like these are, you know, not as good as they are today, but I thought that they would get very powerful, they could definitely draft letters for me, they could certainly maybe help me gather history before visits and maybe look up a guideline for me, so, yeah, started working on tools for myself, and then that ultimately led to Heidi, we raised some money, I took a break in my training and we've kind of gone from starting in Australia, we built really clinician centric tools, and then in the last, say, 18-24 months, we launched globally, and I think by usage we're now one of the most used clinical AI tools in the whole world, so we have more than, I think 500,000 clinicians that have come and used Heidi, and something like, you know, 2 million visits a week, or more than 2 million visits a week now, so yeah, it's kind of insane how far it's gone.
4:54 - MEGAN ANTONELLI: Yeah. Yeah, that's amazing and congratulations. It's, it is crazy. I think that, you know, having been part of and kind of watched the digital transformation within healthcare and to some degree, the resistance to change and the resistance to some of the technologies that have been introduced that it seems like with Ambient technology and this generation of AI tools, the adoption and the potential has just been off the charts. I imagine, you know, you saw similar potential when you founded Heidi Health.
5:31 - TOM KELLY: Yeah, I didn't, I definitely didn't anticipate this speed of adoption, you know, I was used to doctors taking their time and being a bit skeptical of different softwares, you know, before it caught on. But I think I realized now that this is the first time that doctors have ever had a piece of software that actually gives them time back, because the first generation of tools was mostly an exercise of documenting or recording what you were doing, which is incredibly useful, like definitely very useful for everyone else in the system, but for you as an individual, that was another thing I had to do, whereas this actually takes that away to some degree, you know, I can get back to my patient and Heidi can do some of the driving the computer that I would have had to do during the visit. So yeah, it's, and it's pretty simple. You know, you click start, you click stop, that's it, and you get what you need. So for even the most tech averse types, it's not too tricky to experience the value.
6:30 - MEGAN ANTONELLI: Yeah, no, absolutely. And, you know, you spoke a little bit about the, you know, the number of users, the number of clinicians using it, you know, when it comes to kind of those adoption rates and, you know, sort of what's happening versus the reality versus the hype, right, because I think, you know, there's always a hype cycle in healthcare, we know it, and certainly in technology. What are you seeing in terms of those patterns and what have you been seeing over the last, you know, few months, years?
7:01 - TOM KELLY: I think we get a really good snapshot because we have a big US business and it's growing really well, but we obviously came from outside, so we get different countries have different maturity levels, so I would say the US is ahead, a lot of health systems are trying or picking tools, you know, it's the American spirit, productivity, there's like a good incentive for them to be efficient, so they're always seeking to make their systems better where they can. So, it's interesting like in the US we see many systems having mediocre experiences with most of the tools, and I think it's because the kind of brand of product that a CIO would pick isn't necessarily the one that the doctors would really love, and I think we've all experienced it, you know, if you told me before in 2022 that maybe 30-40% of the world would be using a new technology platform to do most of the AI queries, I would have not believed you. But now we look and we see ChatGPT has become like the platform of where this happens. And it's because when such a huge technology shift happens, the whole platform, like everything changes, like everything about the medical record, how you think about how a doctor practices completely changes. And if you're trying to change the world from within the old paradigm, it just constrains you. So what has happened, I think, is that a lot of the tools that medical records are selling to CIOs is really constrained, like it's like a form filler, or it's like a note filler or it's a dictation tool. But we've had these things for a long time, that's not what's exciting about AI tools. What's exciting about AI tools for me and for everyone in the world is that they're completely open plan, you can do whatever you like with it. Sure, as a doctor, getting my notes in easily and single signing on and not having to log in, integration matters, but it's not the thing that makes it great. The thing that makes it great is the product being better, and more flexible and more usable.
9:10 - TOM KELLY: So, I think in the US where we're seeing everyone realizing that this is important and maybe they've spent a bit too much on licenses and not quite getting the adoption they expected and we present a different option of this is still really well integrated, but even in some of our third party reviews and audits and things we're gonna go live with over the next few months, it really shows up, like we're ranked highest by far on clinician usability and the reviews we get are amazing. That actually, in my opinion, should come first over integration because it ends the capability of the product, so it's kind of you need both, but you should prioritize the user experience first. Anyway, so the long story short is I think the revenue and the competition is like almost ahead of the real usage that a lot of systems in the US are seeing, whereas in other markets the other way around. We're seeing incredible adoption, you know, sometimes like half the GPs or primary care PCP doctors of a whole country using Heidi every week, but the systems are still deciding, like how they integrate it or whether they buy it. So yeah, it's interesting, it's like different worlds.
- MEGAN ANTONELLI: No, it is, and I mean, culturally, I think it's so true and it's definitely one of the problems in healthcare in terms of that, just the resistance to change that, you know, they call risk aversion, but enough, it's often more, well, we already sunk so much money into this, we can't quite make the change and there's a little bit also of, you know, nobody lost their job for implementing big guy. And then, you know, when you go, when you do go that risk where you do go the other direction, that tends to the resistance is there. So it's a bit harder. But at the same time, I think to your point that the flexibility and the agility of these tools that are so appealing to the clinician is what makes them, you know, desirable. So if you lock, if you end up locking that out, then you're not going to get the value that you want for sure. So it's really, it is interesting. In terms of where you see that adoption, whether in US or outside, are there certain settings, whether it's the emergency room or you know, primary care, where you're seeing more impact, and better examples or even just notable examples that you can share?
11:47 - TOM KELLY: Yeah, for sure, yeah, I would say, naturally in ambulatory clinics or primary care settings is where, I think it's actually where the clinicians were drowning the most and had the least help, so they were the ones that were the first in most settings to adopt it and just immediately start getting value, you know, they're seeing lots of patients in their day, it's also very, almost like a routine, like the patient comes in a certain way, you stay in the practice all day, it's very structured, so it's kind of easy to adopt a tool that's saving you time. I'd say, although that's the case, for us, we find that actually every specialty stands to gain a lot. So when we've rolled out with let's say Beth Israel, it's across all their specialties out of the gates, so, you know, 34 different specialties, I think we've made more than 1000 templates for them now, and that's, again, that's for us, that's where we differentiate, so the templates and the ability to personalize down to every individual user, because we're not waiting for the access we need from different medical records to do that, we just, we have a great integration too, but we let the users do what they want with Heidi to get efficiency and then when they send it back to the record, we structure it in the way we need to for RCM and everything else that happens next.
13:30 - TOM KELLY: So actually, practically the issue I think is more the current generation of tools that are stuck in the record, do not serve surgeons, they're not useful for carers or nurses because what those clinicians do and where they get the value is, it's like a Venn diagram. The amount of work they do on the record is much less than the PCPs. They have a lot of other stuff, like they're doing operative notes, they're doing consents, they're explaining surgeries to patients, and Heidi's really useful for all those things. But if you're just inside your medical record, then the value to them is not that big, cause in vascular we only used to do like 10 to 20 surgeries. So I have my template, I already know how to dictate, I'm like lightning fast. I don't need some AI trying to make up my notes. So I think it's the, yeah, it depends, like if you're Heidi, actually good for everyone. If you're in the record, probably limited to PCPs and ambulatory generally, and emergency as well is interesting, I think we see quite amazing usage in emergencies. The main challenge there is actually the kind of clinical environment, so, you know, you imagine you're going from room to room, you're seeing, have a few different patients going at once, you need to be able to pause one patient, jump onto the app, pick the next patient, you know, pull through the notes from the previous visit, like it's very mobile kind of heavy setting. So yeah, we have a really good mobile experience and same thing there, like different emergency doctors have different styles, so getting the personalization right is quite important there. But yeah, I think yeah, it comes down to the tool, just depends.
14:58 - MEGAN ANTONELLI: Yeah, no, but that's really interesting because I remember, I, you know, and I'll date myself here, you know, maybe it was Lynn Vogel and maybe I forget if he was, I think it was Intermo a long time ago, like when EHRs were just being implemented and he talked about how, you know, they were implemented and they weren't specialty specific, right? I mean, there was just this one size fits all solution. And you know, that 25 years later, we're sort of still there. And so to talk about that and think about that now we're, you know, in some cases, just layering these AI tools and these ambient tools onto what was already a system that, you know, that brought no one delight. Let's just say that, right? So what you're saying, and I actually didn't know this, but tell us a little bit about what does that personalization look like? Why is Heidi different from some of the others? What does it mean to be able to customize and personalize it so that it feels like it's yours?
16:08 - TOM KELLY: Yeah, so I think, the way, a good analogy because it's obviously, you know, try the product and you can experience it, but I think in a nutshell, most of the tools out there only allow you to fill a certain form. So they may use the word template, but they're still forcing you to put certain content in certain boxes, so it's like, you know, the subjective goes here, the objective goes here, you wanna add a section, no worries, but it has to be like structured in this way. So template is kind of limited. In Heidi, we give you access all the way down to the sentence, so you can say, I want you to start this sentence with the patient presented with, and then insert the rest with the AI please. Like very granular level of specialization. And that sounds maybe like shoulder shrug, like why do I need this, you know, what's the point? It's actually quite important, so for some specialties, like let's say I'm a psychiatrist, the DSM criteria, like the diagnostic criteria will require you to characterize something very precisely, like you have to describe their symptoms in a way that classifies them for a certain diagnosis. So giving people that super granular control means that Heidi can write exactly what you need, not just into the block. Cause you can't really put a sentence in a block, like if I need you to, if the patient has ADHD I need you to put the DSM criteria in, and I need you to write it in this way. Well, how do I do that if it's a form that's for every patient, like I can't build what I need here.
17:50 - TOM KELLY: And so obviously that level of personalization means that we as Heidi have to do a really good job with our customers, so we don't expect every doctor to set this all up. There's lots of other doctors in the team like me who will partner with the organization, understand everything and build them for you. We have like a, I think 10,000 deep library now of different documents that people have created. And then, so that's one kind, and then we also have an always on personalization that learns from edits. So every individual doctor within their own experience of Heidi, any interaction they have, any edit they make, anything they do in the product, Heidi takes as feedback, it's almost like an error. Like if the doctor has to make a change, we failed, they had to use their keyboard, that's a mistake. So then we will slightly adjust their setup to try to avoid that mistake the next time.
18:33 - TOM KELLY: So basically everyone gets a kind of, I guess, vanilla experience in the first instance, and then everyone branches off into different flavors as they edit and as they set up their templates, and it's kind of almost hard to explain because, to non-clinicians, the product looks the same, like when you click stop, it just streams the note. But for the doctors, they know the difference between vanilla or like choc chip, you know, for them it's like the details matter on how they write it cause it's your work, like it's my note. I'm telling colleagues about my patient. It's gonna represent me, so I wanted to make sure it sounds like my notes, it's not some, you know, to use the word of the day, the AI slop, like it's not just some random note. It's my note and written like me.
19:21 - MEGAN ANTONELLI: Yeah, I mean, I would liken it to, I mean, you know, I shouldn't publicly claim loyalty too, but I like Claude over ChatGPT because Claude, I can make my projects very specific and tailored, and they've remembered and they've learned, right? So you, what you're building, what you've got is sort of a learning system that's individualized. And that's to every physician within an organization.
19:43 - TOM KELLY: Correct, correct.
19:44 - MEGAN ANTONELLI: And then?
19:44 - TOM KELLY: And then when we work with the organization side, we have a step that does the CDI so it does the structuring of that document for best billing capture and for integration into the medical record. So although, you know, we might have a urologist who has a section for bladder cancer and a section for the last prostate antigen study, and when it goes into the record, it may get restructured so that it captures billing best and fits into the way the medical record works. So it's still, I think sometimes people view this as like one or the other, like you give flexibility or it's structured. It's not the case, it's more that in order for this to work well with the medical record or for the organization to get any benefit, first, the doctors have to love it. Like they have to use it, they have to get active, every patient, you know, we can talk about creating an AI hospital and Heidi doing care coordination and gap closure and all this stuff in the future, but that only happens if 90% of your doctors are active on the scribe because that's where you get the information, that's where you get the transcript about what to do next.
20:50 - TOM KELLY: And so I think we were talking earlier about the rebrand, so, you know, we, as Heidi, like we're starting to build this care partner that takes the transcript and starts to do many things, not just the notes but, completing tasks for you, generating documents automatically, even phoning the patient a few days later to check in on how symptoms have evolved, and so all these things, I love them because they're very cool, but they still rely on that transcript being there, like the doctor has to have used it, and the only way is if they love the scribe, like you have to go really deep and make it amazing for it to work.
21:29 - MEGAN ANTONELLI: Yeah. And I mean, to some degree, how you were talking about with like the sort of product first versus the integration versus, you know, where we continue to layer technology on a system that wasn't quite built perfectly, right? Which then just makes the system weaker and worse, whereas, you know, what you're doing it seems as you've got this tool that, you know, it's built on the clinician and are there checks and balances for, say, what if that clinician continues to make the same, makes a similar mistake, right? If I spell calendar wrong 800 times will it start to spell it wrong? So are there checks and balances that way to sort of correct on the side of things?
22:20 - TOM KELLY: Yeah, 100%, that's why we call it like a care partner, so it's like, you know, we're not subordinate, like we're a partner, so if, and especially as we're building like an evidence product and like all sorts of things that are coming next. So yeah, of course we need to learn from the user, but we also know what's good for healthcare, what's good for the patient, also what the organization needs, and sometimes they don't match, you know, sometimes, maybe the clinician wants us to write something in a certain way that isn't very good for how billing needs to be captured. And so when we do integrate and work with the organization, we will correct the clinician's notes, we'll make sure that the template's structured in a way where it can work well with everything else. So it's not entirely like everything, you know, to a fault for the clinician, there's still some balance. But I'd say if you're picking a direction to focus on, you still want to skew towards getting them happy and active. And then like pruning the edges to make it work well with the system is fine, that's easy. It's more if you do it the other way around where you just give them like one note type or like very limited anything and it's like, you know, this is AI for you, this is it, like, it's like, oh I thought it was like gonna be like ChatGPT where I was like, no, no, no, like. You know, you click this button in your record and then the notes appear and you're like, I don't really like the notes, and they're like too bad, you know, they're best for billing. That's when you run into issues like you'll only have, you'll still have doctors that love it, like they'll still get glowing reviews, but if you look at the overall group, there might only be 10 or 20% that are really loving it. And there's probably a huge group that are kind of they're like, yeah, it's just not that very good. I'm just using Heidi separately, I don't tell them, you know, so yeah, I think of it in that way, like definitely need checks and balances, definitely need to have like for Heidi, we have to focus on what's great for healthcare, not just for that doctor, but we skew towards serving them more than anything else, in the first instance.
24:26 - MEGAN ANTONELLI: Right. Yeah, no, totally. And I mean, with clinicians in general, I mean, you've talked about the adoption rates being great and sort of that balance of kind of why because it's that personalization, they like the tool. As you, I mean, you know, when you do look across the landscape with different types of physicians or where there is more hesitation, right? I mean, do you have sort of a set of things that you share with your partners to help them get their clinicians in that mindset to sort of ease the adoption and ease that transition?
25:06 - TOM KELLY: Yeah, 100%. We have a really amazing support team now, so basically I think of it as different phases, it's almost like you have to go on the adoption, technology adoption curve within the organization. So you have to find the group within the organization that's already using AI in their own home, at home and in their work in different ways. And so the way we do it is we break it up into different phases, so we have an alpha phase, which for us is actually discovery, so we learn about the specialties, we find out our early adopters and champions, and what we do is we actually get them to create the content for their peers, so we find that, you know, of course, we can make great content, but, if it comes from a peer that they know who's saying that this thing's amazing, then it really helps adoption. It especially works if it's a colleague that they didn't expect who they're like, oh, I didn't know, you know, as if like this person was gonna use it, and they, if they can figure it out, maybe I can too. So I think yeah, generally in that method, and we sort of know internally what are the key things.
26:21 - TOM KELLY: You really wanna protect that first impression. I think once you lose the crowd, it's very hard to undo the perspective, cause there's the, I think the default expectation is that this is gonna be no good, and, you know, they picked a bad product and like they've stuffed it up again, like that's just always the expectation you're going against. So yeah, we, I think we've done a pretty good job. We haven't turned it's got a surprise and delight out of the box because you're dealing with the skeptical crowd as it is.
26:52 - MEGAN ANTONELLI: When you think about, you know, you talked about kind of the AI hospital, the AI partner of the future, the AI care partner, you know, when you think about what does the future look like, and we seem to be moving there very fast in that, you know, as things are getting adopted quickly. What does that look like? You know, what is the future of healthcare look like where clinicians and their AI assistants are working together?
27:21 - TOM KELLY: I think it just looks way more human, cause I think when I say Heidi's doing things on behalf of the clinician, sometimes people imagine, oh so, you know, you're just trying to replace the doctor, you know, the classic thing, and it's like, no, no, no, no. Like Heidi cannot feel a pulse, cannot tell you if a person's sick. I was a surgeon, like surgery is, you can't get anything more examination oriented than that. What I mean is, everything that goes around it, when I explained like why I left to start Heidi, the calling I had was about being that compassionate, knowledgeable person with the patient. Everything that's around that is waste of my time. To be honest, like I, all I wanted to do is read a summary, understand what this person's coming in with, go be with them, be the best doctor in the world, remember their family members, remember what happened last time, talk to them about their day, make a plan of what to do next. Walk out, that plan's being instituted, their visits are being coordinated. If I have questions or I'm uncertain or I don't know the evidence for a certain decision, sure, Heidi's there and I can ask for feedback or I can get clarification in the visit, and maybe if they want, we can even safety net them, like if they're gonna make a dangerous mistake, we say, hey, you should really think about this, or you might have missed this in the medical record. But it's still like the wrapping around me being a doctor, like it's still my responsibility, I'm still in control, I'm still, the patient's still with me, it's just that Heidi's orchestrating a lot, increasing amount of my work outside of my patient care.
29:01 - TOM KELLY: So yeah, I think practically, obviously that sounds, you know, pie in the sky, so practically, I think of it as a couple of key things. So one is evidence, evidence is really important, there's so many clinical decisions that clinicians make, and we have this unique position with them in the visit already, we know who they're seeing, we know the plans they're providing, so, extending from the scribe into starting to give them if they want information about their practice, I think is really exciting, and then doing some care coordination with phone calls, but also text and other modalities is quite interesting. So based on the visit, you know, we know that this person has to go get their blood tests checked and then has to come back after that to talk to the doctor again. Heidi can work in the background to make sure those things happen, so you don't lose patience to follow up. Also for patients that, you know why you have to go do that thing, like why, what do I have to do again, I can't remember the plan, you know, Heidi could share a summary and then text you to say, hey, like, just reminding you, have you gone to get that blood test, is there anything you need, any questions?
30:11 - TOM KELLY: And yeah, I'm really excited about that because, again, like as a doctor, I wish I could have done these things, that would have made me the best doctor ever if I could just call my patients and help them with what they need to do next, but I'm busy so I can't do that. So if Heidi can be an extension of me, then patients get better care and my job is easier and I'm a better doctor, and that's it, like that's how we win, you know, everyone wins.
30:32 - MEGAN ANTONELLI: Yeah, making, you know, and I think that that's always what's gotten me so excited about AI and in its many forms. I mean, technology in the first place, but really what, you know, that change and that shift to how we've been using AI is that it can do all the things that we can't do, that we didn't have time and in a lot of cases. That's the low hanging fruit, you know, and I remember thinking about it in the first time when it was around, you know, we did a lot of stuff around social determinants of health and how we would collect that data and how we would use it, who would use it. And every conversation ended up with no one because no one has time. You know, and now with AI and the applications, there are things we can do. And that processing and the collection process, the assessment, and then the what does it mean and how do we respond to it is all part of that, you know. And I think as you kind of take it from that one example to all of the many examples in healthcare where just that little extra outreach or that little extra touch drives the real engagement, drives the real behavior change, and then you see a real impact, you know.
31:47 - TOM KELLY: Yeah, for gap closures and Medicare and Medicaid, like this is incredible, you know, there's so much information that it's yeah, it's like, you know what you need to do is just who's gonna do the work, like who is there to actually do it. We talked to a lot of organizations that struggle to, not even the doctors, just to find care workers or others that can actually support patients that you know, if they're in a risk contract that are costing them the most, and they know what they need to do, like they need to get them to see their podiatrist, to make sure they've seen the dietitian, to help them with their groceries because their blood sugars are very out of control and, as doctors, we like, if I could sit with that person and coach them, I could do a great job, but it's just really hard, there's just a labor problem of like who actually does the work.
32:35 - MEGAN ANTONELLI: I mean it's time and sort of that ability to practice within your, you know, skill set, your license and all of that. Well, so one of the things in our last few minutes that we always talk about because we have a segment called Five Good Things, and sometimes, I don't know, maybe it's the New Yorker in me, we focus on things that aren't the best things happening in healthcare, although we've taken a more positive view now that we have our segment, but we'd like to close with, you know, what's the bright spot? What is the good thing that you're looking at and seeing as change that's happening, whether it's in technology or healthcare practice in general, that you're really excited about?
33:13 - TOM KELLY: Yeah, I think, I don't know, I'm really excited about getting back to that patient centric model of care, I think, and the specific examples are just the feedback I get, like I saw one literally today that was like, my wife knows that I've been having an affair. That was the message I got, and I was like, oh, what do you mean? And he's like, oh, with Heidi, of course, you know, like my patients say that I, they've never experienced me like this, I'm happier, I'm more friendly, like they're asking like what's going on, and he said he tells all his patients he's having an affair with Heidi. So I think, yeah, so I think cause when I imagined being a doctor, because my doctor actually as a primary care doc didn't actually have a medical record. He used to write things on cue cards and have little dot points and summaries of things. Now, I'm not saying that's better, I'm sure he barely remembered half the things about me, but for me, he used to sit on the heater and he would be opposite from me, and that was why I wanted to become a doctor. It was like that experience with him, it was just like, man, I couldn't imagine being, this being a better job in my life to be like this guy. So, I think that's the model of care we can get back to, you know, Heidi's running in the background doing lots of things, your systems are working for you, not taxing you, and then, yeah, patients just have this delightful experience. You know it's the hardest thing in the world to be sick, no-one wants to be sick, no-one wants to be at the doctor, and to be able to use your knowledge for good and be there with your patient. Every time I get one of those comments funny and delightful it's just the best thing ever, so I think healthcare is gonna get better, not worse, it just takes time.
34:55 - MEGAN ANTONELLI: Oh, I love that. I do have to ask, so why Heidi Health? Why Heidi?
34:59 - TOM KELLY: So it's a bit of a coincidence, so the first AI tool I ever built basically would take your history and give you a differential diagnosis. But you were talking to someone, so it was like, oh, it would be really good if it was someone's name because it would make you feel a bit more comfortable talking about your symptoms. And then someone's like, yeah, we go from history to diagnosis, like HI to DI. And someone wrote on the board, like high dye, and I was like, no one's gonna know how to spell high dye. So then like, yeah, why don't we just call it Heidi? And then, and yeah, the domain was available, so that was it.
35:37 - MEGAN ANTONELLI: Well, thank you so much, Tom. This has been a pleasure, you know, just to learn about kind of your background and what Heidi's doing. And I love hearing about the kind of differences and similarities of the healthcare system across. And it's, you know, sometimes you think that we're so special and unique here in the United States, but in fact, many of our challenges are faced across the globe, right? But thank you again, for those of you who are interested in learning more about Heidi Health, go to the website, Tom, what is the website?
36:12 - TOM KELLY: Heidihealth.com or you can go to Heidi.AI as well.
36:15 - MEGAN ANTONELLI: OK, perfect. And then Tom is on LinkedIn and you can reach him there. We will both be at Health, and I'm sure he'll be at a lot of other events coming up this fall, so be there and we will be in touch.
36:30 - MEGAN ANTONELLI: Thank you so much and thank you all for joining us and this great discussion on how transformational AI can be implemented where it truly counts. This is Megan Antonelli signing off for Digital Health Talks and Health Impact. Have a great day.
Outro: Thank you for joining us on Digital Health Talks, where we explore the intersection of healthcare and technology with leaders who are transforming patient care. This episode was brought to you by our valued program partners Automation Anywhere, revolutionizing healthcare work flows through intelligent automation. Nara, advancing contactless vital signs monitoring. Elite groups delivering strategic healthcare IT solutions. Cello, securing healthcare identity management and access governance. Your engagement helps drive the future of healthcare innovation. Subscribe to Digital Health Talks on your preferred podcast platform. Share these insights with your network and follow us on LinkedIn for exclusive content and updates. Ready to connect with healthcare technology leaders in person? Join us at the next health impact event. Visit Heimpactforum.com for date and registration. Until next time, this is digital Health Talks, where change makers come together to fix healthcare.