How can AI transform the healthcare industry and help solve the challenges of rising costs, staffing shortages, and administrative inefficiencies? In today's episode of Tech Talks Daily, we explore this question with Alex Zekoff, co-founder and CEO of Thoughtful AI.
Thoughtful AI is a company revolutionizing healthcare administration through its innovative, role-based AI agents. These fully human-capable AI solutions—CAM, EVA, and PHIL—are designed to streamline critical but tedious tasks like claims processing, patient eligibility verification, and payment posting.
Healthcare providers in the U.S. currently face mounting pressures, from high staff turnover and costly manual processes to claim denials and prolonged payment cycles. Thoughtful AI's solutions offer a new way forward. By automating these administrative workflows, providers can significantly lower costs, reduce days sales outstanding, and boost claims approval rates. The result? Healthcare organizations can shift their focus and resources from back-office operations to improving patient outcomes.
We discuss how these AI agents combine cutting-edge technologies like RPA, OCR, and large language models to deliver end-to-end automation. With implementation timelines as short as three to six months, these systems not only improve efficiency but also address critical data security concerns by providing full audit trails and real-time analytics without storing sensitive patient data.
Beyond the numbers, we explore the broader impact of AI in healthcare, including how Thoughtful AI's innovations could pave the way for a shift in healthcare spending—reducing administrative overhead and enabling a more patient-centered approach. This episode highlights real-world examples of how organizations are scaling their operations without adding to administrative headcount, achieving tangible results that signal a new era for healthcare efficiency.
What role do you think AI will play in shaping the future of healthcare and government systems? Join us to uncover the possibilities, and share your thoughts!
[00:00:04] Could AI revolutionise the US healthcare system? Today, my guest, co-founder and CEO of Thoughtful AI believes it can.
[00:00:14] They're a company that's tackling some of the most pressing challenges in healthcare administration.
[00:00:19] With staffing shortages, rising costs and the growing complexity of claims processing, Thoughtful AI is leveraging role-based AI agents to transform traditionally manual and error-prone workflows
[00:00:32] into a more seamless, automated process.
[00:00:37] And in this episode today, Alex is going to introduce us all to Thoughtful AI's new solutions from Cam, Eva and Phil,
[00:00:46] new AI agents designed to streamline claims processing, patient eligibility verification and payment posting.
[00:00:54] And by reducing costs, improving efficiency and enabling healthcare providers to focus more on patient care,
[00:01:00] my guest hopes these innovations could mark a turning point for the industry.
[00:01:05] So, big question, can AI hold the key to more efficient and patient-focused healthcare systems?
[00:01:11] Let's explore together as I get my guest onto the podcast now.
[00:01:16] So, a massive warm welcome to the show.
[00:01:19] Can you tell everyone listening a little about who you are and what you do?
[00:01:23] Hi, Neil. Yeah, great to be here.
[00:01:25] My name is Alex Zikoff.
[00:01:27] I'm the co-founder, CEO at Thoughtful AI.
[00:01:30] Started my career 17 years ago building aerospace systems.
[00:01:33] Went to Japan helping Mitsubishi with the company after that.
[00:01:35] After graduating from Berkeley, I got into Gen 1 RPA and AI technologies.
[00:01:40] Started building bots.
[00:01:41] And really thought about a different way of building software that operates software.
[00:01:46] And so, five years ago, kind of got the inspiration to start Thoughtful AI with a mission to democratize this very expensive technology and go into a very what's considered laggard industry like healthcare that's always trying to adopt new technology but seems to be 10 years behind.
[00:02:00] So, really getting access to incredibly complex, expensive technology at a fraction of the price and really focusing on pain points like the revenue cycle management.
[00:02:10] So, fast forward to today.
[00:02:11] We've grown 400% year-over-year growth.
[00:02:14] AI is now the hot talk of the day.
[00:02:17] And we're really excited just to be working with healthcare providers, helping lower their costs, helping deliver better patient outcomes, helping build a really great business along the way.
[00:02:26] And as you said there, AI is the hot topic of the day.
[00:02:30] But you guys were doing this before.
[00:02:31] It was cool.
[00:02:32] So, Thoughtful AI is role-based AI agents.
[00:02:36] For my understanding, you aim to transform healthcare administration.
[00:02:39] You mentioned five years ago you had that inspiration.
[00:02:43] So, I'd love to dig a little bit deeper on that.
[00:02:45] What was it that inspired you to create these AI agents?
[00:02:48] And how have they gone on to solve some of the biggest challenges in the healthcare system?
[00:02:53] Because there's got to be a story there, right?
[00:02:56] Yeah.
[00:02:56] So, it actually goes back to eight years ago when I started building my first bots.
[00:03:00] And I kind of thought, wow, if you think about all of us, this is kind of like we're on the assembly line.
[00:03:07] And so, then if you go back even further in history to 1913, you can think about people actually assembling the Model T in Henry Ford.
[00:03:15] That's what we're doing every day when we log into our terminals and we move data between web browsers and applications.
[00:03:20] And I thought, wow, what happened in that case?
[00:03:23] Well, if you fast forward to the 1980s, if you look at a manufacturing plant for an automobile company, it's all robots.
[00:03:29] And that happened generally around 60, 70 years.
[00:03:32] So, I thought, wow, if you fast forward another 50 years from here, this is all going to be bots and software operating software.
[00:03:38] And then I thought, well, how big is that market?
[00:03:41] There's over $10 trillion of people operating software with repetitive tasks.
[00:03:46] So, then you think, wow, that's wild.
[00:03:48] That's a lot of money just pushing data around.
[00:03:51] One and a half trillion of that is back office healthcare administration.
[00:03:55] Three and a half percent of the US GDP is back office claims processing, including revenue cycle management, administration, putting data into systems.
[00:04:04] And I just think, back then, I thought that was crazy that we're spending that much money processing claims to get paid for providers.
[00:04:13] And then, obviously, that transfers, Neil, to your patient cost, right?
[00:04:18] So, if that bloats there, then that just means we are now the most expensive healthcare system.
[00:04:23] So, I just thought there was a better way of operating software.
[00:04:27] We've built software in this paradigm where we sell software to people, and then we expect people to log into the computer and move data.
[00:04:33] We are now the limiting factor people in software.
[00:04:37] And so, generally, there can be a whole philosophy on what we're going to do if we're not logging into our…
[00:04:43] I've got beliefs in that, and I think it's actually net positive for society not to be behind a computer.
[00:04:48] So, that was the inspiration that I don't think people should be moving data repetitively on computer systems.
[00:04:53] I think we're much better at creative ideas, specifically in healthcare, around doctors and investing more on that side and patient outcomes, less on administrative tasks.
[00:05:02] So, I really just started with that simple idea, and I've kind of just been obsessed with it since and started Thoughtful Eye with a mission to kind of just build a better system.
[00:05:10] So, we have agents…
[00:05:12] We call them agents now.
[00:05:13] Agents that build agents.
[00:05:14] We have bots that build bots.
[00:05:16] We're getting very meta now with the belief that this will all be automated in the next 20 years.
[00:05:21] I love it.
[00:05:22] And if we were to take a look under the hood or sneak behind the curtain of any healthcare institution, things like, as you mentioned there, claims processing, patient eligibility verification, payment posting, they're all traditionally labour-intensive and often error-prone tasks, unfortunately.
[00:05:40] So, how do your Thoughtful AI agents streamline these processes, and what kind of impact have you seen them having on efficiency and accuracy for healthcare providers?
[00:05:51] Because there's a lot of hype around AI, but when we start talking about real meaningful metrics, that's what gets people sitting up and paying attention, I think.
[00:05:59] Totally.
[00:05:59] So, let's kind of baseline the current state of the world today and what could be.
[00:06:03] And so, let's say you're a dental group.
[00:06:05] You've got a bunch of dental locations.
[00:06:07] This is a common sort of customer we would interact with.
[00:06:11] Let's say you've got 300 locations.
[00:06:15] You're a big brand.
[00:06:16] And you're likely going to have around 150 revenue cycle employees collecting money.
[00:06:21] So, that's just 150 FTE just trying to get you paid.
[00:06:24] What does that really translate into?
[00:06:26] Well, we look at a KPI called cost to collect.
[00:06:29] In most cases, when you put things on your credit card or you pay wire, you pay almost zero to maybe 1%, maybe up to 3% on your credit card if it's expensive.
[00:06:38] But generally, you're trying to lower your cost of a transaction.
[00:06:42] In healthcare, that can be up to 5%, 6%, 7% cost to collect.
[00:06:48] You take that and you marry it with denied claims.
[00:06:51] For every $100, sometimes, a provider might only really be collecting $0.80 on the dollar, $0.75 on the dollar.
[00:06:59] Also, they don't get paid right when they have the service.
[00:07:02] We call this day sales outstanding in the industry.
[00:07:05] That can be up to 60, 70 days later they get paid.
[00:07:08] So, they don't even get paid on the service, right?
[00:07:11] It's two months later in that case.
[00:07:13] So, all that meaning it's very expensive to work in healthcare from how the payment rails work.
[00:07:18] Now, let's fast forward to the future and how we're solving that problem.
[00:07:23] We come in and we look at the entire 150 revenue cycle system of people, process, and technology.
[00:07:28] We kind of say, how can we take that call thousand steps, condense it?
[00:07:33] So, usually it's about deletion first, about process deletion.
[00:07:37] And then we say, hey, we can hook it all up.
[00:07:39] We can install things like Ava, which is an Elibilogy agent that can process unlimited verifications versus being limited by what a person can do per day.
[00:07:49] Or claims processing.
[00:07:51] Or on the back end would be payment posting.
[00:07:53] Once they get the claim back, posting the remit to the bank.
[00:07:56] You take all that together and you can lower all those KPIs I talked about.
[00:08:00] So, cost to collect, we can get down to less than 1%, which is unheard of.
[00:08:05] Like, not even possible really in the human world, but in the agent world it is possible.
[00:08:10] And for a CFO, that means maybe 4% going to their bottom line.
[00:08:14] 4% on $300 million, an extra $12 million a year.
[00:08:18] That could be going to doctors, front end staff, you know, more research.
[00:08:23] All of that money could be going to better patient outcomes.
[00:08:26] Then we think about day sales outstanding.
[00:08:28] Right now, 60 days, we can get that less than 30 days.
[00:08:30] So, an extra month, you collect your cash earlier and then denied claims rate.
[00:08:35] Let's say you're for every dollar you send out as a claim, you only get paid $0.90 on the dollar.
[00:08:41] 10% denials rate.
[00:08:42] We can help that lower that 10% denials to 2%.
[00:08:47] So, collecting $0.98 up front.
[00:08:50] Or we focus on clean claims rates.
[00:08:52] So, clean claims out the door.
[00:08:53] Error-free claims.
[00:08:55] Coming back and getting paid without getting denied.
[00:08:57] So, all of that to say, it's more money for the provider.
[00:09:01] It's faster money.
[00:09:02] It's just a 10x improvement on their current way they're getting paid.
[00:09:06] And when we're thinking about this industry, other big areas of problems or challenges are staffing shortages, rising admin costs, all major concerns for any healthcare provider.
[00:09:17] So, again, how do you see automation with AI agents helping alleviate some of those challenges and ultimately allow providers to focus more on those patient outcomes?
[00:09:29] I mean, you mentioned some pretty big financial numbers there.
[00:09:32] But anything else you can share?
[00:09:34] Yeah.
[00:09:34] So, in this industry, turnover is a huge problem, especially in revenue cycle.
[00:09:39] Up to 40% turnover per year.
[00:09:41] So, again, on that 100 people, you might be turning over 40 people a year.
[00:09:45] Extremely costly.
[00:09:46] And, again, I'm not even adding in the cost of turnover when I'm talking to you conservatively about how much the providers have to hire, train, make revenue cycles, employees productive.
[00:09:56] And then we also have to think about the group of revenue cycle employees right now, the sort of age range is starting to retire.
[00:10:04] And we don't have Gen Z millennials coming to backfill those jobs.
[00:10:08] There's 12 million people in the U.S. operating healthcare administrative tasks right now.
[00:10:12] That expected number demand based on healthcare demand needs to be 18 million people.
[00:10:17] There's not 6 million people backfilling those jobs with the current turnover rates.
[00:10:21] So, this is going to end up being a mission critical solution in the next 10 years.
[00:10:26] One, our healthcare costs continue to balloon per capita.
[00:10:29] And our actual average life expectancy has been going down the last 10 years.
[00:10:34] And that's all driven by cost flow.
[00:10:36] That's all driven by bureaucracy.
[00:10:37] That's all driven by what I'm explaining to you now.
[00:10:39] It's just it's complicated.
[00:10:41] And there's no incentive really for people to go into healthcare.
[00:10:44] So, if you think about talent, the smartest people, where do they go?
[00:10:47] Tech.
[00:10:48] Banking.
[00:10:49] They get paid, right?
[00:10:50] Like, that's how capitalism works.
[00:10:52] And I just think there are some really perverse incentives when you think about how the provider
[00:10:58] payer mix works.
[00:11:01] And not to bring up a hot topic, but obviously the shooting of the United Healthcare CEO,
[00:11:05] very tragic.
[00:11:06] But that's highlighting people are now seeing the denials rates by insurance providers.
[00:11:11] It's now becoming a very hot topic about, you know, not really highlighted that because
[00:11:16] I think they published a chart that United Healthcare was denying.
[00:11:19] They had the highest denial rates, 30%.
[00:11:21] And then, you know, other people have.
[00:11:23] But that's it's highlighting to the population that we have a problem, right?
[00:11:28] There's weird incentives and it's causing real issues.
[00:11:32] People not getting services for healthcare and dying really at the end of the day.
[00:11:36] And so, I think in a country like ours where we're so rich that we have a very bifuricated
[00:11:42] system right now.
[00:11:43] And I think generally, you know, we're trying to solve that.
[00:11:47] And we're trying to solve that, you know, not only be technology, but, you know, our mission
[00:11:51] is a little broader than that.
[00:11:52] It's to fix the U.S. healthcare system is our main mission.
[00:11:54] We think payments is first, but there's a lot of other ways we're trying to get in and
[00:11:58] help providers, you know, solve these complex challenges.
[00:12:02] And you briefly mentioned claim denials, cost to collect issues, and days outstanding or
[00:12:08] DSO there.
[00:12:09] So, I've got to ask just to dig a little bit deeper on that.
[00:12:12] How is it that Thoughtful AI solutions specifically address these financial pain points?
[00:12:18] Because I suspect there'll be many in the industry or many healthcare providers that are looking
[00:12:22] to improve these areas and improve their revenue cycle.
[00:12:25] How do you do this with ThoughtAo?
[00:12:27] So, we are now what we've considered level three AI.
[00:12:30] They're called agents.
[00:12:31] So, when I started, we had level on bots.
[00:12:34] Bots are deterministic, they're hard-coded, they're brittle.
[00:12:37] Meaning, you code them, they can work through a screen path like a person.
[00:12:42] But if they come into something they haven't worked through before, they break.
[00:12:46] And a person's smarter than that.
[00:12:47] A person can see that the application changed where a button is and then click submit.
[00:12:51] So, we started in bots and then OpenAI, DropChat, Chippie Team, models came out.
[00:12:58] And then when you combine bots plus models plus other technologies, you can get to level three
[00:13:03] AI, which is agents.
[00:13:05] Agents is a full replication.
[00:13:07] Let's say, Neil, you're a revenue cycle employee.
[00:13:09] And you walk in, you've got 200 clicks or steps you do every day across eight systems.
[00:13:14] There's a lot that happens that you don't realize that you're doing.
[00:13:18] And there's a lot of technologies that would go into replicating Neil's day-to-day work and end.
[00:13:23] There's optical character recognition, which would be sort of your eyes.
[00:13:27] There's the large language model, which is sort of your intelligence.
[00:13:30] We use a little bit of RPA for, you know, repetitive task.
[00:13:34] And all of that gets combined together in what's called an agent.
[00:13:37] So, Ava is an AI agent, fully human capable.
[00:13:40] It takes anywhere between three to six months to train her.
[00:13:43] And then we install that into the revenue cycle system.
[00:13:48] And then now that Ava is running the nWorkflow, not a human.
[00:13:53] And we help train people to run Ava, to manage it.
[00:13:56] Ava spits off all of the data live streaming as she's working in all the systems.
[00:14:01] There's exception reports.
[00:14:02] There's insights.
[00:14:03] There's actually insights into the revenue leakage now.
[00:14:06] And so, now revenue cycle employees can go monitor the analytics and help improve the system, not just operate the system.
[00:14:13] And so, that's how we're actually being able to lower all of the key metrics I had described earlier.
[00:14:19] It's because the revenue cycle employees now, instead of just clicking buttons, they're actually helping improve the revenue cycle system by working with Ava, Cam, Phil, Paula together.
[00:14:28] And we call that a system of agents.
[00:14:30] So, we will build Ava, deploy.
[00:14:34] Now Ava runs eligibility.
[00:14:35] We'll build Cam, deploy.
[00:14:37] Now Cam runs claims processing.
[00:14:38] Same thing with Phil.
[00:14:39] Phil runs payment posting.
[00:14:41] So, you don't need 100 people running the system anymore.
[00:14:44] You might need 20, 30.
[00:14:46] Those 70 people get typically either reallocated up to patient experience, which is like customer support, technical support, or patient support, or managing the agents.
[00:14:56] So, we have a front end where they can manage the execution of the agent, the exception reports.
[00:15:02] They can help train future agents.
[00:15:03] We're actually kind of seeing an initiative now where providers are getting with us AI first.
[00:15:09] So, they're like, wow, not only can we automate revenue cycle, what about our HR processes, IT?
[00:15:15] We built our system so we can automate any role in the world on our platform.
[00:15:20] And so, we name the agents for the ones that we can do the fastest, most predictably, Ava, Cam, Phil, Paula, Dan, Cody.
[00:15:26] But really, any SOP standard operating procedure can be turned into an AI agent on our platform.
[00:15:31] And what would you say to anybody that's listening, thinking about that, thinking, wow, that is incredibly cool, but, oh man, I could lose my job.
[00:15:37] Do you see a natural shedding of human employees, or do the healthcare providers that you're working with, are they upskilling or moving them around the business?
[00:15:49] What do you say there?
[00:15:50] So, again, I try to look at history to see, because this happened before, and this has happened many times, history has proven that technology will come through and sort of automate sort of jobs that might end up being repetitive.
[00:16:03] And then what happens after that?
[00:16:04] So, let's use telephone operators back in the day.
[00:16:07] Telephone operators, very similar to revenue cycle.
[00:16:09] There's a lot of, you know, literally manually switching and putting in cords.
[00:16:13] Technology came and automated that.
[00:16:15] So, what happened to telephone operators?
[00:16:17] I did some research on this.
[00:16:18] They went into technical support and customer support jobs a decade later.
[00:16:22] And then I looked at the average compensation for those same pool of operators, telephone operators.
[00:16:28] They were making, you know, back then $3,500 was a lot, $3,500 a year.
[00:16:32] And in a decade, when they went into the higher value jobs, the more technical support, customer support, they were making close to $7,500.
[00:16:40] So, adjusted for inflation, they actually had real income increase.
[00:16:45] And so, I think the same thing is happening.
[00:16:47] We're going to need to reskill for the modern economy.
[00:16:50] I'm sure higher ed is not going to say this.
[00:16:52] I've been, you know, I went to master's and get all that.
[00:16:54] But I think we're going to have to rely less on the education system for skills.
[00:17:00] And we're going to have to rely more on YouTube, if I'm honest.
[00:17:02] And we're going to have to get people into skills and trades and out into plumbing and electrician.
[00:17:06] And again, working with patients and not just on computers.
[00:17:10] I think we've all gotten a little addicted to the easiness of the logging in the computer, having all the apps.
[00:17:16] It's very nice if we can work from home.
[00:17:18] But I think generally there's going to be a big shift into physical world jobs over the next 20, 30 years.
[00:17:23] I think that's a net positive.
[00:17:24] I think we see a lot of decaying infrastructure in the country.
[00:17:27] I think we see a lot of opportunities for skill-based work.
[00:17:31] And so, I think there's going to be, I think whichever entrepreneur wants to take on this task will be another,
[00:17:36] this will be a trillion dollar opportunity.
[00:17:37] Basically, re-skilling the people that will get reallocated from AI into higher value jobs,
[00:17:44] that is a multi-trillion dollar opportunity.
[00:17:47] Because people will re-skill.
[00:17:49] People are resilient.
[00:17:51] They'll go on YouTube and learn how to, you know, do this.
[00:17:54] And then they'll get paid more.
[00:17:56] And I think that's just generally we need to encourage that incentive and not kind of focus on like what happens to everyone.
[00:18:01] It's like, all right, well, we do have a lot of problems in the world.
[00:18:04] So, those first principles, let's reallocate people to the highest order problems.
[00:18:08] Let's make sure there's incentive where they get paid more than they're making now.
[00:18:12] And then kind of help transition those people into those new skill-based jobs.
[00:18:18] Such a great point.
[00:18:20] And thinking out loud, the adoption of AI in healthcare could also be met with skepticism due to concerns around things like,
[00:18:27] again, just thinking off the top of my head, reliability, human oversight, maybe security of data as well in AI models.
[00:18:34] So, how are your thoughtful AI agents ensuring its agents perform at a high level while also maintaining things like trust and transparency with providers, etc.?
[00:18:46] So, two things we will be, we're very transparent on.
[00:18:49] We don't store any PHI data.
[00:18:51] We use serverless technology.
[00:18:53] So, if you've seen the movie Men in Black, there's like this little thing that can erase memory.
[00:18:58] So, every time the AI agent runs, yeah, it forgets everything.
[00:19:03] We can't even get into the container when it's running.
[00:19:05] So, it's like completely secure.
[00:19:06] There's no storage of data, which is really important.
[00:19:09] We only store things that we're told to store metadata on the runs.
[00:19:13] And that is one really important part of the aspect of the build.
[00:19:17] Two would be, we don't throw data in models.
[00:19:20] We use models to operate agents.
[00:19:22] So, it's a little bit different of a shift in how we think about agents.
[00:19:25] So, we don't have a model that's just throwing in all this like operational data and saying solve the problem.
[00:19:31] We're using things like computer use models, which would be, it's called agentic browser navigation.
[00:19:37] So, instead of having to code, hey, Neil's logging into an EHR application and performing a task, we can prompt our agent to log in, do the task, take the data out, and then come and put it into a system.
[00:19:51] That's not even code anymore.
[00:19:53] That's just a computer use model.
[00:19:54] But we're not putting any data or healthcare data into that model.
[00:19:59] It's more of an operational computer use model.
[00:20:01] So, again, no data is being put into the model for security reasons.
[00:20:05] And so, we really focus on that.
[00:20:07] As far as oversight goes, we have a front-end system.
[00:20:10] We call it Empower.
[00:20:11] And that is really all tracking of the agents.
[00:20:14] Every single click action an agent runs versus a human is tracked.
[00:20:18] And there's a report.
[00:20:20] You can audit it.
[00:20:21] You have every single thing that an agent was doing and this clean output.
[00:20:27] So, then if there's any case of like, was it the agent that messed up?
[00:20:31] Agents never mess up.
[00:20:33] They just run through the task.
[00:20:34] But you have that report.
[00:20:36] And that really helps ensure safety and trust on both fronts so that you know there's not a rogue agent going on just operating your system and gone off the rails.
[00:20:45] And we talked about the challenges in healthcare, some of the big problems that you're on a mission to solve there.
[00:20:52] And if we were to fast forward a few years into the future, how do you envision that future of AI in healthcare administration?
[00:20:59] And are there any other areas of the industry that could maybe benefit from a similar role-based AI solution?
[00:21:05] Anything that excites you around that in the future?
[00:21:09] If I fast forward 10 years, I think agents...
[00:21:13] Well, I think the two industries that are going to benefit most, one, healthcare, very inefficient system, cost bloat.
[00:21:18] You got to follow the money.
[00:21:19] Cost bloat will be where AI goes.
[00:21:23] So, healthcare will completely look different.
[00:21:25] I truly believe healthcare costs are going to come down.
[00:21:27] We're going to have a much more efficient system.
[00:21:30] We're going to see better patient outcomes.
[00:21:31] And we're going to see the trend line go in the right direction, not the wrong direction.
[00:21:35] And I think AI is going to be an enabling technology for that, both on the administrative side, but also on the clinical side.
[00:21:41] And, you know, helping with diagnosis and actually like co-pilots for doctors and how they help solve, you know, different ailments and things like that.
[00:21:50] The other thing that we are very interested, obviously, of big inefficient systems government.
[00:21:55] And we are in Austin, Texas and already working with the state of Texas on, you know, really sound AI policy, as well as deploying AI potentially in, you know, their regulated environments.
[00:22:08] What we really build is we're able to deploy AI in highly regulated environments.
[00:22:12] And so, government is another really exciting area where the talk of the town is cost bloat, the doge, right?
[00:22:18] We've got D-Lot and Vivek coming out here and saying they're going to slash and burn.
[00:22:23] And I think, you know, thoughtful stances, we do need to reduce costs, but no pun intended thoughtfully.
[00:22:28] So, that's what we want to do is get into think about government and role-based AI there and what are the roles that we can create.
[00:22:35] The next, you know, whatever we're going to call it, Ava Camphill, we'll create new named agents for those roles and find out where there's some government inefficiency that we can scale and help reduce costs there.
[00:22:44] So, we're really excited by those two industries, healthcare and government.
[00:22:47] And we've talked a lot today around the role of AI and automation and how it actually leads to better patient outcomes.
[00:22:55] And there's a lot of hype around AI and a lot of people struggle to understand the actual real-world impact that it can have.
[00:23:01] So, are you able to share any examples of how your technology has ultimately enabled a healthcare provider to allocate more resources to patient care and less administrative burdens to them?
[00:23:12] Just to leave everyone on a positive note and maybe help them visualize how it would work in their world too.
[00:23:18] Yeah.
[00:23:19] So, I'll use one customer example here.
[00:23:23] They are a dental group.
[00:23:25] They had basically a group of revenue cycle employees servicing five dental locations and they were going to have to hire 20 people.
[00:23:33] They're scaling fast.
[00:23:34] So, they were hiring 20 people to basically scale revenue cycle.
[00:23:39] Fast forward to the day, they've installed AI agents.
[00:23:41] They started two years ago.
[00:23:42] They were very early adopters of this technology.
[00:23:44] And now the same group is not servicing five locations.
[00:23:47] It's servicing 25 locations.
[00:23:49] They were still going to hire those 20 people, but instead of hiring those 20 people to collect money, they hired those 20 people on the front end to serve patients.
[00:23:57] And so, that's usually exactly how it works with when you're going to allocate your budget every year.
[00:24:03] Are you going to allocate to administration or clinical?
[00:24:06] More of those dollars are going to clinical.
[00:24:07] And that's what we want to see.
[00:24:08] We want to see the ratio go from basically it's about 60% clinical, 40% admin.
[00:24:14] We want to see that shift over the next decade to 80%, 90% clinical, 10% admin.
[00:24:21] So, it's really about dollar reallocation, getting people up more on the patient side, which is going to lead to better patient outcomes.
[00:24:27] I think that's a perfect moment to end on.
[00:24:29] So, thank you so much for sharing your insights today.
[00:24:32] But before I let you go, I'm going to ask you to leave one final gift for everyone listening.
[00:24:36] We have an Amazon wish list where I ask my guests to recommend a book that they check out.
[00:24:41] So, what book would you like to add to our list and why?
[00:24:45] One of my favorite books I read 20 years ago, it was actually on Bill Gates' list.
[00:24:50] And I kind of was like, what's Bill Gates' favorite books?
[00:24:53] Moonwalking with Einstein.
[00:24:55] It is an amazing book about memory.
[00:24:57] We actually are putting memory into our ages now.
[00:25:00] And I kind of think about that full circle.
[00:25:03] My big takeaway about that book is the perception of your life is based on the novelty you create.
[00:25:09] So, if you every day are like traveling and doing new things and, you know, really creating a non-repetitive life, your life will feel really long.
[00:25:17] And so, like it might feel like you led six lifetimes.
[00:25:20] But if you have, you know, every day you log in, you do the same thing, you log out, you eat the same meal, you will feel like your life's really fast.
[00:25:27] And so, I kind of 20 years ago in my early 20s when I read that, I was like, well, then I'm just going to like make sure every week is different.
[00:25:35] I travel a lot.
[00:25:36] There's no repetitive part.
[00:25:38] So, I just want to live, I want to feel like I've lived a really long life.
[00:25:41] And I think like that book's all about memory and memory palaces.
[00:25:45] It's really interesting about the human mind.
[00:25:47] So, I'd recommend anyone who's just interested in that memory or how we, because we're losing our memory with computers.
[00:25:54] Like we're becoming less reliant on our brains.
[00:25:56] And I think actually that's a critical skill to remember and have memory.
[00:26:01] So, I think anyone who's really interested in memory, it's a really great book to kind of, a reminder how to remember things.
[00:26:08] Wow.
[00:26:09] I need to check that one out myself.
[00:26:11] So, I'll be not only adding it to the wish, I will be looking at that.
[00:26:14] And for anyone listening wanting to find out more information about thoughtful AI, maybe digging a little bit deeper on anything we talked about or even contacting your team,
[00:26:22] where would you like to point everyone listening?
[00:26:25] Yeah, our website usually thoughtful.ai.
[00:26:28] Our LinkedIn, you can find us thoughtful AI.
[00:26:30] You can email me directly, alex.thoughtful.ai if you have questions or want to talk about AI.
[00:26:36] And yeah, we're happy to chat.
[00:26:38] Looking forward to, you know, again, going deeper into healthcare and deploying some pretty amazing AI solutions here in the future.
[00:26:44] One of the things I've loved about talking with you today is how traditionally healthcare providers have been forced to do those tedious but critical work tasks that usually are costly, manually intensive and error-prone processes.
[00:26:57] And when we're looking at AI and automation to solve some of these challenges and enable providers to stop hiring for those roles but benefit significantly from higher productivity, that is all incredibly cool.
[00:27:10] But there's a great line used towards the end of our episode there about talking about shifting those or the allocation of those dollars that we're going to admin tasks to patient tasks, getting people in front of patients, working with patients.
[00:27:23] That's got to be a huge step in the right direction.
[00:27:26] So thank you for sharing your story with me today.
[00:27:29] Really appreciate your time.
[00:27:31] Awesome.
[00:27:31] Thank you, Neil.
[00:27:32] Appreciate it.
[00:27:33] I think my guest has shown today that the integration of AI and healthcare administration does have the potential to not only reduce costs and efficiencies but also refocus energy and resources to what truly matters.
[00:27:46] We're talking about improving patient outcomes here.
[00:27:49] You don't want somebody sat behind a desk endlessly scrolling.
[00:27:53] Surely it makes more sense to have more people with patients.
[00:27:55] And by automating tedious processes with role-based AI agents like Cam, Eva and Phil, Thoughtful AI is attempting to pave the way for a smarter, more sustainable healthcare system.
[00:28:07] So a big thank you to Alex for sharing his vision and the transformative impact that AI is having on healthcare and its potential to do even more.
[00:28:15] But over to everybody listening, how do you see AI reshaping industries like healthcare or indeed others that you might be listening and be involved in?
[00:28:25] Something you don't associate with tech or AI.
[00:28:28] I'd love to continue the conversation with you.
[00:28:30] Yes, you.
[00:28:31] I know you listen every day.
[00:28:32] I don't want you to just be passively listening.
[00:28:35] I'd love to sit down with you two and hear your thoughts.
[00:28:38] So whether that be joining me on a podcast episode or sliding into my DMs, either is fine with me.
[00:28:45] LinkedIn, Instagram, X, just at Neil C. Hughes.
[00:28:49] My email, techblogrideratoutlook.com.
[00:28:52] I'm the easiest guy in the world to find, so why not send me a message?
[00:28:55] So keep those messages, questions, requests to come on the show coming over, and we can regroup again tomorrow morning.
[00:29:02] But that's it for today.
[00:29:03] So bye for now.
[00:29:05] Bye for now.
[00:29:05] Bye for now.