Latest / HumanifAI HealthTech / The Healthcare AI Trends Defining 2026 w/ Dr. Geeta Nayyar
Transcript
- 0:00Francesca: 2025 was the year that healthcare AI flooded the market with pilots, press releases and big predictions.
- 0:08Francesca: At the recent JP Morgan conference in 2026, the storyline was heard and with loud and clear its time for AI solutions to prove real impact or step aside.
- 0:20Francesca: Now today, we're breaking down the trends in the healthcare AI space that will define success in the year ahead.
- 0:34Francesca: Welcome back to the Humanify Health Tech podcast, where we explore how AI is rewriting the rules of healthcare and putting people right back at the center where they belong.
- 0:46Francesca: I'm your host, Francesca Vera, head of Marketing at Net Health.
- 0:50Francesca: And if 2025 was the year of healthcare AI where it exploded onto the scene, 2025 is the year it has to prove itself.
- 0:59Francesca: Healthcare systems are moving past pilots and the headlines are starting to ask tougher questions about what we value.
- 1:06Francesca: Safety, governance, and real clinical impact.
- 1:10Francesca: Now, joining me today is Dr. Geeta Nayar, physician technologist, author, and occasional co-host of the show.
- 1:18Francesca: And she's gonna help me break down the trends that are poised to separate the wheat from the chaff in the healthcare AI space, and help us decide what leaders should be paying attention to in 2026.
- 1:30Francesca: Dr. G, welcome back.
- 1:31Francesca: Jumping right in, AI scribes have become one of the fastest adopted healthcare AI use cases, right?
- 1:39Francesca: This is autonomously capturing clinical conversations and generating those notes real time.
- 1:44Francesca: Now in 2026.
- 1:46Francesca: This next evolution seems to be quite clear.
- 1:49Francesca: Those tools are moving beyond just transcribing, much more towards anticipating and suggesting, and some of the next steps that clinicians might even ask.
- 1:59Francesca: I also noticed I was doing some extra reading Dr. G and I saw in some recent interviews in the UCSF magazine with Dr. Bob Wachter.
- 2:08Francesca: I hope I'm pronouncing his name correctly.
- 2:10Francesca: He can reach out afterwards if I mess that up.
- 2:12Francesca: But he cautioned that AI scribes become more seamless and physicians may start to, to require, um, being a little less diligent about verifying what the systems produce and that might actually increase.
- 2:26Francesca: Some of the risks that error starts slipping into medical records.
- 2:30Francesca: So my question to you, Dr. G, is how C can clinicians really stay vigilant and ensure that accuracy across every stage of that care journey is taken into account rather than just gradually outsourcing the judgment to the system?
- 2:45Dr. Geeta: Sure.
- 2:45Dr. Geeta: Well, first of all, thank you so much for having me back, Francesca.
- 2:47Dr. Geeta: It's always a pleasure to be on and, and thank you for that.
- 2:51Dr. Geeta: So, I love this question because the, the idea that, uh, AI scribes are gonna take over the note is perhaps very exciting when we think about documentation and contribution is con contribution to physician burnout.
- 3:04Dr. Geeta: But I think you're right.
- 3:06Dr. Geeta: It, it really is important that we don't fall asleep at the wheel, right?
- 3:10Dr. Geeta: So.
- 3:10Dr. Geeta: I think it's important that for clinicians, we think of this very much as like when we had a medical student or a resident.
- 3:17Dr. Geeta: So when we were doctors in training long before the EHR, and even as we transitioned to the EHR, for the most part, medical students or residents, if you saw patients with them, they wrote the note.
- 3:27Dr. Geeta: They wrote the note.
- 3:28Dr. Geeta: They were essentially a human, a human scribe, long before we had scribes.
- 3:32Dr. Geeta: And it was the same because they were not, they did not have the authority to have the notes signed.
- 3:37Dr. Geeta: The system would not accept a signature from someone who was not the attending.
- 3:41Dr. Geeta: So I think of this as very similar, that those safeguards have to be in place and that we're really careful that we're not allowing the description, the, the ambient, uh, scribe service to actually become the author.
- 3:54Francesca: I think that's such a smart point, and I love the fact that you bring in the entire enterprise journey.
- 3:59Francesca: Um, so as we shift into our next trend, which is really becoming a favorite theme for us on this podcast, um, interoperability, right?
- 4:06Francesca: So interoperability has long been framed as this notion of just.
- 4:10Francesca: Pure data exchange and it being a little bit of an issue, particularly as AI is transforming data exchange from just mechanical movement, um, into more intelligent interpretation.
- 4:22Francesca: Right?
- 4:22Francesca: And we know that systems can certainly automatically reconcile medication lists.
- 4:27Francesca: Um, across various settings, they can identify gaps during transitions and they can translate disparate data formats into so that this unified clinical kind of narrative.
- 4:38Francesca: Um, now to bring this to light, I was actually chatting to a wound care director just last month who told me something that really ties into this point really well, Dr.
- 4:47Francesca: G, and I'd really like to hear your point in it.
- 4:49Francesca: She said.
- 4:50Francesca: We don't really have an interoperability problem anymore.
- 4:53Francesca: We really have a translation problem.
- 4:55Francesca: Um, and her example was, she says, I've got a, the skilled nursing facility that gets data from the hospital, but it arrives in this like wall of text that nobody has any time to read.
- 5:05Francesca: And she says, what I need is not more data.
- 5:09Francesca: What I need is I need the system to tell me that this patient has a. Stage three, pressure in injury, right?
- 5:15Francesca: That debridement occurred on day two, that there's an antimicrobial protocol that they're on, and here's a red flag to watch for in the next, say, 48 hours, like that shift, moving from just records, right and into surfacing more clinical meaning at the moment it matters, is really what she's looking for.
- 5:35Francesca: So my question to you, Dr. G, is how does AI change interoperability from just simply moving healthcare care records from one system to the other into generating that clinical meaningful insight and data that clinicians can actually use during these healthcare transitions?
- 5:53Dr. Geeta: So I really love that point, and I love that story because we really are, this isn't about just getting information anymore.
- 5:59Dr. Geeta: This is about using AI to interpret the in information, right?
- 6:03Dr. Geeta: Just like, just like that colleague of yours mentioned, right?
- 6:06Dr. Geeta: What are the medications that the patient is actually on?
- 6:09Dr. Geeta: What are the care gaps that are actually missing?
- 6:12Dr. Geeta: And what did I just miss, right?
- 6:14Dr. Geeta: What did I miss?
- 6:14Dr. Geeta: What's the red flag that will really make a difference in care?
- 6:18Dr. Geeta: So I think it's really about that intelligence layer and really interpreting the data.
- 6:23Dr. Geeta: The simple data exchange, thankfully, is not really where we're at anymore, but it's really about what does it mean clinically and how and what should I be alerted to?
- 6:32Dr. Geeta: And I think the only piece of this that we have to be mindful of as AI helps us interpret data.
- 6:38Dr. Geeta: Sometimes it makes it sound really good, Francesca, right?
- 6:41Dr. Geeta: And it's a great hallucination.
- 6:43Dr. Geeta: So that's the only caveat is that sometimes it makes everything seem so coherent and so perfect when in fact something is actually missing.
- 6:51Dr. Geeta: And so again, that accountability, that clinical oversight is always, always gonna be needed.
- 6:56Francesca: Totally fair point.
- 6:58Francesca: And I think you actually said that there as well.
- 7:00Francesca: You know, AI is learning from those documentation patterns, um, and it's learning to start to predict some of those missing elements, right?
- 7:07Francesca: Those flagging those potential denials before submission, ensuring that there is some, you know, specific quality measure in place.
- 7:15Francesca: Um, and in fact, um, according to the American Hospital Association.
- 7:19Francesca: Hospitals, and this was a mind blowing fact that I read.
- 7:22Francesca: They spent over $39 billion annually on administrative compliance activities.
- 7:28Francesca: Meanwhile, you've got cybersecurity and compliance leaders increasingly citing AI tools for real-time monitoring and audit readiness.
- 7:36Francesca: So the question here really starts to become, how could predictive AI really reshape compliance in 2026 by identifying documentation gaps this time or denial risks before claims are even submitted rather than after the audits occur.
- 7:52Francesca: So it's a shift from the clinical use case into the back of office documentation compliance.
- 7:58Francesca: Base.
- 7:59Dr. Geeta: Absolutely.
- 8:00Dr. Geeta: And look, we spend billions of dollars trying to correct errors after the fact.
- 8:06Dr. Geeta: What the opportunity is here with, with AI is that we're able to actually prevent, right?
- 8:12Dr. Geeta: The best compliance strategy is a strategy that involves prevention, right?
- 8:16Dr. Geeta: It shouldn't be after the fact, but if it can, in fact, find the coding mistakes, find the denial triggers, find, you know, where are the special specialty specific measures missing.
- 8:27Dr. Geeta: In real time so that it's not actually submitted incorrectly, but actually proactive.
- 8:32Dr. Geeta: So I think it's terrific.
- 8:33Dr. Geeta: I think it's one of those amazing use cases for AI that could really change the game.
- 8:38Dr. Geeta: And really the enterprise should be looking at this as a priority.
- 8:41Dr. Geeta: 'cause it's also, I think, one of the simpler implementations of ai.
- 8:45Francesca: I think it's such a smart point.
- 8:47Francesca: Um, alright, here's another trend that I would love to get your perspective on.
- 8:51Francesca: So, um, AI is analyzing all the patient type of engagement patterns, right?
- 8:56Francesca: And that's both in HEB and in remote therapeutic monitoring.
- 9:00Francesca: And that's really just.
- 9:01Francesca: Start to diagnose, uh, any adherence risk before the patient actually starts to disengage.
- 9:06Francesca: So think of like your at home compliance or treatment plans, particularly say in like a rehab therapy scenario, but also candidly in wound care.
- 9:14Francesca: Now the system starts to flag who needs intervention and then starts to recommend whether there's some sort of a need for a motivational check-in or some clinical assessment or even some program simplification that might be needed.
- 9:27Francesca: I'm thinking of like some of the senior folks that we care for.
- 9:30Francesca: Maybe the exercise is too hard or difficult for them.
- 9:34Francesca: So in these specialty care settings, like a rehab therapy or a wound care scenario, how can AI move past that patient adherence from a retrospective metric?
- 9:44Francesca: You were just talking about retrospectively shifting into more of this proactive intervention strategy.
- 9:50Francesca: What have you seen and what do you recommend?
- 9:52Dr. Geeta: The same thing, right?
- 9:53Dr. Geeta: It's, it's not really helpful when you find out a patient has disengaged after the fact.
- 9:59Dr. Geeta: If AI can actually help us shift that and actually prescribe interventions that will work in real time, whether it's a diGeetal check-in.
- 10:07Dr. Geeta: Whether it's motivational outreach, whether it's a clinical reassessment, but really kind of gauging the mood, right?
- 10:12Dr. Geeta: You and I both have have kids, right?
- 10:15Dr. Geeta: We know when our kids are sort of, their mood is off, they're not receptive, and we change our strategy as a parent and we have to do the same thing clinically.
- 10:22Dr. Geeta: We have to be able to identify when a patient is.
- 10:26Dr. Geeta: On their way out and starting to slip from the system and starting to disengage, missing an appointment, missing a diGeetal check-in.
- 10:33Dr. Geeta: And so again, I think this is one of the best use cases for AI is to actually get the patient and, and get them at the right time when they're receptive as well as when they need it.
- 10:43Dr. Geeta: And so.
- 10:44Dr. Geeta: I think about this as, you know, we've always thought of adherence as a report, right?
- 10:48Dr. Geeta: This is how many people didn't do this or didn't do that, but I think it should be an early warning system, right?
- 10:53Dr. Geeta: Mayday, mayday, this is where we need help.
- 10:56Dr. Geeta: This is where this person is at this moment, and, and swoop in there and get them at the right time in the care journey.
- 11:02Dr. Geeta: I mean, when we think about wound care, right?
- 11:03Dr. Geeta: We don't want the stage four wound.
- 11:05Dr. Geeta: We don't want it to get there.
- 11:07Dr. Geeta: We want them early.
- 11:08Dr. Geeta: And so we also know when we have that warning system, it makes a difference in clinical outcomes.
- 11:14Francesca: Uh, you know, I think you just threw out a really great idea for a teenage app monitoring app, fa heater.
- 11:20Francesca: I might, I might reach out to you after the show.
- 11:23Francesca: We can kind of, uh, see even get some seed money going.
- 11:26Francesca: No, but all in all seriousness, I think it's, it's really a key point.
- 11:30Francesca: Um, you know, we do a lot with data and that proactive intervention, really helping patients get a complete, um, sort of, you know, clinical experience and really being able to deliver better outcomes as a result of that proactive approach, I think is really a game changer for, uh, where AI is going and the application in 2026 and beyond.
- 11:51Francesca: Um, okay.
- 11:51Francesca: Shifting gears a little bit now to something a little bit more, uh, practical and, and speaking about something that.
- 11:58Francesca: Um, has been embedded when we're talking about apps has been embedded into some of these specialty workflows.
- 12:03Francesca: So computer vision as an example, we've embedded a lot of that into specialty wound care flows, for example, um, in wound care, really specifically as we're tracking smart wounds using smartphones, um, assessing movement qualities, say during a rehab therapy.
- 12:20Francesca: Scenario, uh, even detecting if we've got pressure, injury risks or other sort of applications and using AI to really capture that clinical reality that was previously undocumented or required, you know, a manual way of measuring whether that was a, a paper ruler or just relying on a subjective measurement or, uh, a note.
- 12:40Francesca: Now we also know that.
- 12:41Francesca: In America itself, chronic wounds actually affect roughly 6.5 million patients.
- 12:47Francesca: I, I mean, it's mind blowing when you think about that.
- 12:49Francesca: And so the obvious application for an AI based computer vision application is really increasingly being turned to in order to track progression from injuries.
- 13:00Francesca: Um, one particular, uh, example of a, of a use case that I came across just recently was in wound care.
- 13:07Francesca: We're seeing that really happening at scale, and we're seeing that we're starting to move from this research novelty of just standard care into tracking those healing trajectories.
- 13:17Francesca: Kind of like the proactive care you were just talking about, Dr.
- 13:19Francesca: G. So measuring tissue changes, flagging deterioration risk, but.
- 13:24Francesca: Here's the catch.
- 13:25Francesca: Specialty care doesn't have the luxury of being just good enough.
- 13:29Francesca: From an AI perspective, when you're documenting not just for the patient, but you're also looking for like for medical care audits, or if you're trying to defend something from a medical necessity, that visual ai.
- 13:43Francesca: Needs to be audit grade, accurate and not just directionally helpful.
- 13:47Francesca: Now, in your experience, I think you've come across this a couple of times, as computer vision starts to enter clinical workflows, what safeguards do you think we need to ensure that image-based AI interpretations leverage in order to remain clinically viable?
- 14:03Dr. Geeta: You bet.
- 14:03Dr. Geeta: So again, I think this is, again, an awesome opportunity with ai, but we have to keep in mind that there has to be that, that human authority, right?
- 14:12Dr. Geeta: Because AI can measure the pixels, right?
- 14:15Dr. Geeta: But it can't measure the meaning, right?
- 14:17Dr. Geeta: What is the, the meaning of the granulation?
- 14:19Dr. Geeta: What is the meaning of the wound depth and, and what does that actually mean?
- 14:23Dr. Geeta: And so I think it's a, it's a terrific tool, but it is a really good example of needing that.
- 14:28Dr. Geeta: Override by the human.
- 14:30Dr. Geeta: And, and I say that very much also as a rheumatologist who deals with, with skin diseases all the time.
- 14:36Dr. Geeta: You know, there's variations in skin, there's variations in devices, there's variations in the way that the data is being picked up by the device.
- 14:44Dr. Geeta: And so the safeguards have to be around all of those things.
- 14:47Dr. Geeta: Diverse training, data sets, clinical validation, clear confidence scoring, human override.
- 14:53Dr. Geeta: And really that's what it comes down to is the, is the machines are only gonna pick up.
- 14:57Dr. Geeta: The pixels, but the clinician is really picking up what that means.
- 15:01Dr. Geeta: This is time for this intervention on the wound.
- 15:03Dr. Geeta: Uh, or it's not.
- 15:04Dr. Geeta: And so I think really those are the safeguards that we're looking for.
- 15:08Francesca: I remember you telling us a story, and maybe you wanna just repeat that for your, for the audience, but you were talking about on an X-ray seeing one version of the treatment and you know, an AI tool might just go in one direction, but after you asked a couple of really pointed questions, you were able to get a completely different sort of diagnosis.
- 15:25Francesca: Can you tell our readers a little bit about that, uh, story?
- 15:28Dr. Geeta: Sure.
- 15:29Dr. Geeta: So, you know, uh, we've all had our, our use cases with, with ai and one of them was basically a patient that.
- 15:37Dr. Geeta: Their records.
- 15:38Dr. Geeta: Everything in their records said non-smoker, right?
- 15:40Dr. Geeta: Everything in the record said this.
- 15:41Dr. Geeta: This individual was not a smoker, but when I looked at the x-ray, it was clearly a patient with COPD, clearly a chronic smoker.
- 15:49Dr. Geeta: And when I consulted the, the patient to just check on, you know, why is the entire chart for the past 10 years saying that, you know, he's not a smoker.
- 15:56Dr. Geeta: He said, oh yeah, well I did quit, right?
- 15:58Dr. Geeta: I did quit a few years ago, and basically what happened, Francesca, is the note was being copied and pasted.
- 16:04Dr. Geeta: The note had just been copied and pasted, and so when you look at the report from AI on that x-ray, COPD did not, did not even come up because based on the data that was given to ai, which was erroneous.
- 16:17Dr. Geeta: It wouldn't have.
- 16:18Dr. Geeta: Right?
- 16:18Dr. Geeta: It wouldn't have.
- 16:19Dr. Geeta: And so it came to all of these really rare diseases, and that's where we have to be really careful because actually AI was correct with the data it was presented, but using human judgment, human experience, I knew that that had to be impossible.
- 16:32Dr. Geeta: Right.
- 16:32Dr. Geeta: And then I went looking kind of for where the disconnect was.
- 16:35Dr. Geeta: And so again, these things just can't be on autopilot.
- 16:38Dr. Geeta: It's, it's a great tool, like anything, but the tool is one part of the picture, one part of the diagnosis, and one part of the treatment plan.
- 16:46Francesca: So important.
- 16:47Francesca: And I think what you're saying there is, you know, even a specialized AI tool would have, um, a little bit more, you know, sort of knowledge and depth and yes, you still need that human in the loop.
- 16:58Francesca: I don't disagree, but I was chatting with a rehab therapy director who's been in the field for like 20 plus years, um, just the other day.
- 17:05Francesca: And she gave this example, which I think will bring a smile to your face.
- 17:08Francesca: She was saying things like.
- 17:10Francesca: Um, she was testing for just one of the new foundational model tools, AI tools out there, and she asked her to just write a, just a basic progress note for her patient's recovery, right from a, from a stroke.
- 17:20Francesca: And the output looked really great.
- 17:22Francesca: It was professional, it was well written, grammar was all correct.
- 17:26Francesca: It was completely wrong.
- 17:28Francesca: And she said what it was doing is it was recommending a range of motion and exercises for someone that just would not be able to, you know, to do that considering the condition and use terminology that would've gotten flagged in a Medicare audit.
- 17:42Francesca: Um, and her reaction was just so perfect.
- 17:44Francesca: She said, this is why you can.
- 17:45Francesca: Plug into generic ai and I was thinking of you specifically during this conversation.
- 17:50Francesca: She said Specialty care has context and workflow nuance is so important and clinical judgment, right?
- 17:56Francesca: Those patterns that you just can't learn from the open internet are vital.
- 18:01Francesca: And so I think the question really starts to become whether those foundational models that we're using today continue to get smarter.
- 18:08Francesca: We know they will.
- 18:09Francesca: And it's, it's sort of whether they'll actually start to understand specialty specific care.
- 18:15Francesca: And will those then start to train those unique workflows that we're all, you know, working through.
- 18:20Francesca: So if AI starts to really understand these functional outcomes, then we can actually start to, uh, predict perhaps that.
- 18:28Francesca: Outcomes, delivery, quality measures, et cetera, will be different because those models are off operating at a different level.
- 18:35Francesca: It's not just standard model.
- 18:36Francesca: It's now special specialized AI plus the AI in the in the loop.
- 18:41Francesca: So as a foundation, right, if we start talking about these models and make this assumption that they're gonna continue to get more capable, how do you see the balance evolving between these broad.
- 18:51Francesca: General purpose AI tools, right?
- 18:53Francesca: And some of these generic EHRs and generic tools into more highly specialized clinical AI systems that are built for specific workflows.
- 19:02Francesca: What do you think about that and what's your guidance?
- 19:05Dr. Geeta: Sure.
- 19:05Dr. Geeta: Well, first of all, I, I agree, right?
- 19:07Dr. Geeta: I mean, the foundations are great, but healthcare is complicated and certainly specialty care is complicated.
- 19:12Dr. Geeta: There is no one, one size fits all.
- 19:15Dr. Geeta: So general foundational models.
- 19:17Dr. Geeta: Are gonna be great for general things, but specialty care requires nuance.
- 19:21Dr. Geeta: And so I think that the best is hybrid, right?
- 19:23Dr. Geeta: You have the foundational model, but really the nuances of specialty care are going to have much more value, much more value to the clinician, much more value to the care team, much more value to the patient.
- 19:34Dr. Geeta: So I think that needs to be the goal, but I also think.
- 19:37Dr. Geeta: The fact that we're going in that direction is incredibly exciting.
- 19:41Dr. Geeta: Again, as a rheumatologist dealing with many skin diseases, many wounds, it's exciting because we have a lot of nuance, right?
- 19:47Dr. Geeta: We have a lot of nuance in lupus rashes, lupus patients, and a general AI tool is just not gonna get me there, right?
- 19:54Dr. Geeta: It's not gonna get me there.
- 19:55Dr. Geeta: And so understanding that and embracing that and making those models even more intelligent is, is really exciting.
- 20:01Dr. Geeta: Again, as a tool, not.
- 20:03Dr. Geeta: Not driving the Not driving the plane.
- 20:05Dr. Geeta: Right.
- 20:06Dr. Geeta: And so just being mindful of that, again as a tool, but a better tool.
- 20:08Dr. Geeta: A better tool for sure.
- 20:10Francesca: So, um, if we talk about like enterprise AI and how that's gonna start to depend on federated trust.
- 20:17Francesca: If we think about the fact that intelligence really lives where that data is created, right?
- 20:21Francesca: Gi you talking about understanding and knowing where the data's coming from.
- 20:26Francesca: Um, and if we think about that as those sources being things like specialty EMRs or diGeetal health tools and point of care systems, uh, when we start to think about federated learning and how that allows.
- 20:38Francesca: AI to improve from different collective patterns and also just the ingestion of new data without moving into sensitive PA patient data.
- 20:46Francesca: 'cause we know that's, of course, you know off limits.
- 20:49Francesca: What does that really mean for enterprise level AI data readiness and what does that actually look like for some of the systems as we move into 2026?
- 20:58Francesca: How should leaders really think about structuring that governance so that innovation can continue to move quickly but without creates creating any unnecessary friction.
- 21:08Dr. Geeta: So look, you don't scale technology.
- 21:11Dr. Geeta: You scale trust infrastructure, right?
- 21:13Dr. Geeta: This is about building trust, understanding that there's no technology that's perfect, including ai.
- 21:18Dr. Geeta: And so for me, this is about governance models.
- 21:21Dr. Geeta: This is about trust, what I call trust infrastructure and ensuring accountability.
- 21:26Dr. Geeta: 'cause that's really what it's about.
- 21:27Dr. Geeta: So that looks like cross-functional ai.
- 21:30Dr. Geeta: Governance councils, that looks like defined model validations.
- 21:33Dr. Geeta: Continuous performance monitoring.
- 21:35Dr. Geeta: Monitoring, clear clinician feedback, right?
- 21:38Dr. Geeta: And again, how you implement how and where is the human in the loop.
- 21:42Dr. Geeta: But really those kinds of measures, those kind of cross checks is really assuring the accountability and that's where we have to continue to evolve and get smart.
- 21:52Francesca: No, it's a, it's a huge point.
- 21:54Francesca: Um, I do think this is the year where we continue to move from a lot of those pilots into actual proof, and I think that's gonna be a key area that folks are going to take a look at.
- 22:04Francesca: But now also shifting into a slightly different a piece.
- 22:08Francesca: Not only do you need all those models, you also need workforce readiness.
- 22:12Francesca: Right.
- 22:12Francesca: So let's talk a little bit about how AI literacy is really gonna define what workforce readiness looks like.
- 22:18Francesca: Um, we know that AI absorbs this documentation burden, right?
- 22:22Francesca: Like that is a huge thing that it's helped with us.
- 22:24Francesca: It's also helping with routine scheduling and supply chain coordination so that therapists can really focus on things that matter, right?
- 22:31Francesca: Really focusing on, on the complex reasoning, focusing on the patient motivation, and ultimately hands, hands-on treatment, really working at the top of their license.
- 22:41Francesca: So my question for you, Gida.
- 22:43Francesca: As AI becomes more embedded into these clinical and administrative workflows, how should we have healthcare organizations think about re-skilling?
- 22:52Francesca: You know, there's AI literacy and then there's redefining roles, and then in technology that augments the professional rather than disrupting their day-to-day.
- 23:01Francesca: What do you recommend?
- 23:02Dr. Geeta: So first of all, I couldn't agree with you more, right?
- 23:05Dr. Geeta: AI is not meant to replace clinical reasoning.
- 23:08Dr. Geeta: I actually think it's there to elevate, and to your point, if we're not worried about documentation scheduling, we can actually.
- 23:16Dr. Geeta: Practice at the top of our license, build a relationship, use our judgment, motivate our patients, educate our patients.
- 23:23Dr. Geeta: So all really, really good stuff.
- 23:25Dr. Geeta: I think when it comes to workforce readiness, look, it's a new technology.
- 23:28Dr. Geeta: It's a new, new tool.
- 23:29Dr. Geeta: So what are its limits?
- 23:31Dr. Geeta: Making sure the workforce understands what are the limits of the technology?
- 23:35Dr. Geeta: Where do they need to be mindful and be analyzing when they're, they're also not sure and not taking it as gospel.
- 23:42Dr. Geeta: Uh, knowing when to override, knowing that you can override, right?
- 23:45Dr. Geeta: That's also critical that you are there for a reason.
- 23:48Dr. Geeta: You're not there to just watch the technology, right?
- 23:51Dr. Geeta: You're there to ensure accuracy and ensure that there's no harm done as well as well to patients.
- 23:57Dr. Geeta: Um.
- 23:58Dr. Geeta: Interpreting confidence scores and recognizing bias patterns.
- 24:01Dr. Geeta: I think this is a big one, and speci specifically when we think about health equity, right?
- 24:05Dr. Geeta: Even the technology has bias because it's a manmade product, right?
- 24:08Dr. Geeta: With manmade, uh, data that it's interpreting.
- 24:11Dr. Geeta: And so all of those things I think are really critical when we think about how we get this workforce ready to use the technology.
- 24:18Francesca: And for those that are listening today, if you had to give them one tip on how to embrace ai, what might that be?
- 24:25Dr. Geeta: Oh gosh.
- 24:26Dr. Geeta: I would say embrace it.
- 24:27Dr. Geeta: Definitely don't fight it.
- 24:28Dr. Geeta: It's here to stay.
- 24:29Dr. Geeta: I think it's gonna make you better and happier in your patients too.
- 24:32Dr. Geeta: But I also think it's really important that we understand what the limits are of these technologies and give the feedback.
- 24:40Dr. Geeta: 'cause we are in a state in the industry where we're still figuring this out and so don't think it's fully baked.
- 24:45Dr. Geeta: Don't think, oh, this is the way it's gonna be.
- 24:48Dr. Geeta: This, no, give the feedback.
- 24:50Dr. Geeta: If you're on the front lines, you're seeing patients.
- 24:52Dr. Geeta: Something didn't go right with that wound.
- 24:53Dr. Geeta: Something that you saw was obvious, the technology didn't.
- 24:56Dr. Geeta: You've gotta give that feedback, you've gotta give that feedback.
- 24:58Dr. Geeta: You'll make the organization better.
- 25:00Dr. Geeta: You'll, you'll find yourself also being really more that much more excited about the innovation and your role in it.
- 25:05Francesca: So that what absolute sage advice, Dr. Uh, Dr. G, you sort of recapped it all there.
- 25:11Francesca: The accountability, the involvement, the inclusion, the act of participation in AI and embracing it.
- 25:17Francesca: Um, and I think that feels like a really proper place for us to wrap our discussion.
- 25:21Francesca: Although between you and me, Dr.
- 25:21Francesca: G, I could chat to you forever.
- 25:24Francesca: Uh, really hit on some.
- 25:26Francesca: Top trends today.
- 25:27Francesca: Really appreciate that.
- 25:29Francesca: We're gonna go ahead and let Dr. G go because she's actually on her world tour and is off to, uh, meet with another couple of, uh, you know, key executives, especially on the topics of ai.
- 25:40Francesca: Um, it was wonderful to chat to you today.
- 25:42Francesca: Thank you again for joining us and as always, we appreciate your insight and look forward to having you back.
- 25:47Dr. Geeta: Thank you so much for having me, Francesca.
- 25:49Dr. Geeta: Look forward to seeing you in real life soon.
- 25:51Francesca: Absolutely.
- 25:57Francesca: So some key takeaways from today's conversation.
- 25:59Francesca: First, AI in 2026 is shifting from this reactive tool to much more embedded intelligence, and whether that's in the ambient systems that we spoke about.
- 26:09Francesca: Or interoperability layers that we were chatting about, or even predictive compliance.
- 26:14Francesca: The focus is moving forward and towards surfacing insight at the moment of care rather than after the fact.
- 26:20Francesca: That's a key point.
- 26:22Francesca: Secondly, specialization really does matter.
- 26:25Francesca: So from computer vision that we discussed in wound care to adherence prediction in say a rehab therapy environment, AI is becoming deeply trained in the nuance of specialty workflows, not just general purpose automation.
- 26:39Francesca: And finally, readiness is all about the people as much as it is about the platforms.
- 26:45Francesca: You heard Geeta saying, for instance, be accountable, speak up, be involved, governance, ai, literacy, and all these federated learning models.
- 26:54Francesca: And that's really what human AI collaboration is about, and that's what will determine whether organizations start to unlock real value.
- 27:02Francesca: Or just simply layer on more complexity onto an existing system.
- 27:06Francesca: And one pattern we're actually watching really closely at Net Health is that the companies that are winning in AI aren't necessarily the ones with the fanciest models.
- 27:15Francesca: They are the ones with the deepest workflow visibility.
- 27:19Francesca: So when your platform sits.
- 27:20Francesca: Already inside the clinical decision making process, that's when documentation care transitions compliant, workflows, for example, become much, much more proficient and much more effective because you're not just bolting AI onto the outside, you're really embedding intelligence exactly where it matters most.
- 27:40Francesca: And that's a wrap for this episode.
- 27:42Francesca: Thank you to all of our listeners out there for tuning in.
- 27:46Francesca: And if you're joining us from Spotify or YouTube, we'd love to hear your thoughts and comments below.
- 27:51Francesca: And of course, don't forget to leave us a rating while you're there.
- 27:55Francesca: Once again, I'm Francesca Eb, head of marketing at Net Health, and I'll see you next time.