3047: The Future of Healthcare IT: Humanizing Data with AI at Notable Systems

[00:00:03] [SPEAKER_01]: Welcome back to the Tech Talks Daily Podcast. Now, we talk a lot about AI on this podcast, but I recently found myself thinking bigger and asking the question, what if AI could do more than just automate processes? What if it could truly humanize healthcare and transform the vast quantities of data that is handled?

[00:00:25] [SPEAKER_01]: Well, today I'm going to be joined by Steve Johnson, the CEO of Notable Systems, a company that's at the cutting edge of AI-driven healthcare document processing. With a background in both technology and healthcare, Steve and his team are pioneering a platform that significantly reduces the time and effort required for data entry, making healthcare not just more efficient, but more patient-centric.

[00:00:52] [SPEAKER_01]: And the fact that Notable Systems has just secured an 8.8 million in Series A funding, a testament to their growing recognition of the potential to revolutionize healthcare IT. But even more than that, before we even talk about Notable, my guest today has got a truly inspiring backstory, which I urge you to listen to as well.

[00:01:16] [SPEAKER_01]: And we'll also explore how this balance of AI and human insight is key to improving patient satisfaction, reducing insurance claim denials and ultimately making healthcare more responsive. So if you're interested in learning how AI is reshaping the future of healthcare and what this means for both providers and patients, you're going to love this one.

[00:01:38] [SPEAKER_01]: Are you ready to make 2025 the year that your voice gets heard? Whether you're an entrepreneur, business leader, or just someone with a story to tell, I've got three podcasting options to help you kickstart the new year with your own podcast.

[00:01:54] [SPEAKER_01]: Option one, you record and let me, your host, Neil C. Hughes, do the rest. Simply hop on a call, record your interviews, and leave absolutely everything else to me.

[00:02:05] [SPEAKER_01]: For just $2,000, I'm going to edit, produce, and publish a season of 12 episodes, making sure they go live on Spotify, Apple Podcasts, and beyond. It's simple as that.

[00:02:17] [SPEAKER_01]: So if you or your business are interested in getting a podcast live and you want to work directly with me, visit my website, techblogwriter.co.uk, email me, techblogwriter at outlook.com, or send me a DM on any social platform such as LinkedIn, X, or Instagram. I'm just at Neil C. Hughes on every single platform.

[00:02:37] [SPEAKER_01]: Option two, if you want more than just editing and publishing, maybe you're a large enterprise that needs support around website management for your podcast, promotional clips, engagement, promotional campaigns.

[00:02:49] [SPEAKER_01]: Well, I've partnered with AIpodcast.ing, and they can help you with all of that. And best of all, if you use the code Neil C. Hughes, they'll give you 25% off your first month.

[00:03:02] [SPEAKER_01]: Option three, if you are more of a DIY enthusiast, Libsyn. This is the podcast host that I host this show on. And if you use promo code TBW, that will get you up to two months of free podcast hosting services.

[00:03:19] [SPEAKER_01]: So for example, if you entered that code on November the 2nd, you won't pay a dime until January the 1st.

[00:03:25] [SPEAKER_01]: So no excuses, whether you want full support, a bit of help, or you want to go it alone, I've got something for everyone.

[00:03:32] [SPEAKER_01]: So let's get your podcast rolling for 2025.

[00:03:37] [SPEAKER_01]: But now it's time for me to take you on a trip all the way to the US, where my guest is waiting to speak with us today.

[00:03:43] [SPEAKER_01]: So a massive warm welcome to the show. Can you tell everyone listening a little about who you are and what you do?

[00:03:51] [SPEAKER_00]: So I'm Steve Johnson, founder and CEO of a new company called Notable Systems.

[00:03:57] [SPEAKER_00]: And we are an AI-based intelligent document management system company.

[00:04:05] [SPEAKER_00]: And we are focused on the healthcare space today.

[00:04:08] [SPEAKER_00]: And what we do is we work for mostly companies that sell medical products like knee braces and insulin pumps and wheelchairs who need to be paid by insurance companies.

[00:04:23] [SPEAKER_00]: And in order to be paid by insurance companies, they have to collect all sorts of paperwork from the patient, from the patient's physician or therapist.

[00:04:32] [SPEAKER_00]: And all that paperwork comes in to the medical equipment company.

[00:04:37] [SPEAKER_00]: And up until about now, all that paper had to be processed by hand.

[00:04:42] [SPEAKER_00]: So the pages would come in and that's of all shapes and sizes and no standardization.

[00:04:48] [SPEAKER_00]: And a swap of pages would come in.

[00:04:51] [SPEAKER_00]: And today, people have to look at pages on one screen and type into another screen.

[00:04:57] [SPEAKER_00]: So it's a very manual process of finding information in the pages in order to create a claim that goes to the insurance company.

[00:05:04] [SPEAKER_00]: My company, Notable Systems, has a whole slew of techniques, many of them AI techniques, that finds the information in these pages automatically.

[00:05:15] [SPEAKER_00]: And it is very fast, very accurate, and very, very helpful to the people who are doing this job in getting their job done faster.

[00:05:27] [SPEAKER_00]: There is always going to be a need for some human supervision to make sure the data is absolutely accurate because no AI system is flawless.

[00:05:36] [SPEAKER_00]: And so we've got great tools for helping our customers check the things that need to be checked.

[00:05:41] [SPEAKER_00]: So our system is good at knowing what it doesn't know and directing the staff to pay attention to the things they need to pay attention to.

[00:05:49] [SPEAKER_00]: So we've been in business for about six years.

[00:05:52] [SPEAKER_00]: But really, the first five of those years have been in development mode with a few anchor customers in the medical space.

[00:06:00] [SPEAKER_00]: And the last year, we've come to market, and I'd say in all modesty, we're really tearing it up.

[00:06:06] [SPEAKER_00]: So I think we're poised to own the medical equipment space in terms of data extraction and intelligent document processing.

[00:06:16] [SPEAKER_01]: And you really are tearing it up.

[00:06:18] [SPEAKER_01]: That's one of the things that put you on my radar and why I was excited to get you on the podcast today.

[00:06:23] [SPEAKER_01]: And I am looking forward to learning more about Notable Systems and the great work you're doing there.

[00:06:28] [SPEAKER_01]: But before we do, one of the things I always try and do with my guests is learn a little bit more about their origin story.

[00:06:35] [SPEAKER_01]: What put them on this path?

[00:06:36] [SPEAKER_01]: And when I did a little research on you, you've got quite an origin story, too, that even then began, I believe, with you selling a company to America online.

[00:06:46] [SPEAKER_01]: For anyone listening, hearing about you for the first time, can you tell me a little bit more about that earlier stage in your career?

[00:06:53] [SPEAKER_00]: So way, way back, seems like an eternity ago, because the Internet really didn't exist, or at least the web didn't exist.

[00:07:01] [SPEAKER_00]: The public-facing Internet didn't exist back in the early 90s.

[00:07:05] [SPEAKER_00]: I was a graduate student at Harvard, and a professor friend of mine said George Lucas had just come to campus, the creator of Star Wars.

[00:07:15] [SPEAKER_00]: And he told these engineering professors they needed to think about digital image compression.

[00:07:22] [SPEAKER_00]: And that was because cinemas would, in the future, begin to receive films over telephone wires digitally instead of in tin canisters like they had always done.

[00:07:34] [SPEAKER_00]: And so the professor described this to me, and instead of finishing my dissertation, I started tinkering with digital image compression and ended up with an algorithm that made it possible to send pictures over telephone lines.

[00:07:51] [SPEAKER_00]: Back in the early days when people were dialed into these online services on these squealing modems at very slow data speeds, these online services like America Online and Prodigy and CompuServe were the big ones.

[00:08:04] [SPEAKER_00]: And up until about 93, when I finished the algorithm, it was too hard, too slow to send pictures over telephone lines, over these dial-up modems.

[00:08:15] [SPEAKER_00]: And so my algorithm made it possible to send pictures in seconds.

[00:08:19] [SPEAKER_00]: And so Steve Case at America Online heard about me and my algorithm.

[00:08:24] [SPEAKER_00]: I was a guy in a garage with my best friend keeping the lights on.

[00:08:27] [SPEAKER_00]: There were two of us in the company.

[00:08:29] [SPEAKER_00]: He ended up licensing the technology so that AOL became the first with pictures in 1994.

[00:08:36] [SPEAKER_00]: And then he ended up buying my company when there were about 50 engineers a couple of years later.

[00:08:42] [SPEAKER_00]: And then I was running most of his technology division for the rest of the 90s, which is a really good place for a geeky guy like me to sit on the planet, which was really AOL, the front door to the internet for most people, at least in the United States.

[00:08:57] [SPEAKER_00]: And then it had grown to Europe by then as well.

[00:09:00] [SPEAKER_01]: Oh, what a fantastic story.

[00:09:03] [SPEAKER_01]: I remember those days very well.

[00:09:04] [SPEAKER_01]: The sound of a US robotics modem connecting.

[00:09:08] [SPEAKER_01]: It's something of wonder.

[00:09:10] [SPEAKER_01]: You've seen so much change since then, haven't you?

[00:09:13] [SPEAKER_01]: So I've got to ask, how did you go from there to where we are in 2024?

[00:09:18] [SPEAKER_01]: You've got notable systems here founded by a group of technology and healthcare experts with decades of experience between them, building successful organizations and inventing technology that makes life better and work easier.

[00:09:33] [SPEAKER_01]: What happened there?

[00:09:34] [SPEAKER_01]: I feel there's got to be a story there too.

[00:09:37] [SPEAKER_00]: Yeah.

[00:09:38] [SPEAKER_00]: Yeah.

[00:09:38] [SPEAKER_00]: So when I left AOL about 2000, I've started various other companies.

[00:09:43] [SPEAKER_00]: So in the recommendation engine space and the targeted advertising space, and I was living in Orange County, California at the time.

[00:09:52] [SPEAKER_00]: This doctor friend of mine who had just built a orthopedic hospital called Hoag Orthopedic Institute in Irvine, California, came to me and said, hey, brand new facility, but the IT systems are still from the 70s.

[00:10:05] [SPEAKER_00]: And that's because of these closed systems made by Epic and Sterner, these other big legacy IT companies.

[00:10:12] [SPEAKER_00]: And he wanted me to come in and see if there was any way I can modernize the way information was captured about what happens to a patient when they're in a hospital or a clinic.

[00:10:22] [SPEAKER_00]: And I brought some of my old colleagues from the AOL days into one of the clinics of Hoag.

[00:10:30] [SPEAKER_00]: And the clinic was still on paper.

[00:10:32] [SPEAKER_00]: And I learned, and this was around 2017 or so, that most clinics in America like paper as an input device.

[00:10:42] [SPEAKER_00]: It enables them to maintain eye contact with the patient.

[00:10:46] [SPEAKER_00]: You can get a lot of information on a single page.

[00:10:49] [SPEAKER_00]: You don't have to scroll around.

[00:10:50] [SPEAKER_00]: And so paper was still really the medical device, the information device of choice.

[00:10:57] [SPEAKER_00]: But they wouldn't have minded finding the information digitally.

[00:11:02] [SPEAKER_00]: So fill out the paper form.

[00:11:04] [SPEAKER_00]: And then if there's some way to turn it into a digital record that you can access on a computer, that would be great.

[00:11:10] [SPEAKER_00]: And so we actually built a system, the very first step Notable Systems took, was it created a way for these clinics to fill out all of their paper patient records during the day and then scan everything in.

[00:11:23] [SPEAKER_00]: And then within, well, by the time they came back in in the morning, all of their paper records would have been turned into digital records that they could access on screen, fully audited.

[00:11:34] [SPEAKER_00]: So it would tell them where there were missing fields or missing signatures or something was illegible or didn't check out, you know, was a bad birthday, for instance.

[00:11:47] [SPEAKER_00]: And it saved them a whole lot of trouble because they didn't have to page through everything and check everything manually.

[00:11:53] [SPEAKER_00]: And so when we finished that system, and it required all kinds of things, handwriting recognition and developing a way to find meaning on a page.

[00:12:03] [SPEAKER_00]: So there were all sorts of different forms and different free flowing rows that they would write clinical notes, for instance.

[00:12:12] [SPEAKER_00]: We had to glean information from all of that.

[00:12:14] [SPEAKER_00]: So we had to develop all sorts of techniques that could do that.

[00:12:17] [SPEAKER_00]: And we realized when we finished that, that the clinical market was too small.

[00:12:22] [SPEAKER_00]: We had something very general and very powerful.

[00:12:25] [SPEAKER_00]: And so we went and found a couple of other, we kind of thought of them as sandbox customers, customers who could help us continue to develop the technology.

[00:12:35] [SPEAKER_00]: And we found one in the financial services arena called Creststone.

[00:12:39] [SPEAKER_00]: And we found another in the orthopedic business called DJO.

[00:12:43] [SPEAKER_00]: It's now called Anovus.

[00:12:44] [SPEAKER_00]: And they put us to work on other use cases.

[00:12:48] [SPEAKER_00]: In the financial case, it was statements would come in to these wealth management companies.

[00:12:55] [SPEAKER_00]: And there were data elements in the financial statements that needed to be put into their reporting system.

[00:13:02] [SPEAKER_00]: And that was always done by hand because every report's a different format.

[00:13:05] [SPEAKER_00]: Well, we automated that completely.

[00:13:07] [SPEAKER_00]: So statements would come in and we fully automated and cross-checked.

[00:13:12] [SPEAKER_00]: And so it was very, very validated information that would go from the statements into their reporting systems.

[00:13:17] [SPEAKER_00]: And then the orthopedic company, they had a thousand sales reps out in the field selling things like knee braces and spine braces, things like that.

[00:13:26] [SPEAKER_00]: And so in order to get these products paid for by the insurance company, a claim would have to be put together that required information coming from the doctor or the therapist.

[00:13:37] [SPEAKER_00]: And that information would come on demographic sheets, on clinical notes, on prescription pages.

[00:13:45] [SPEAKER_00]: And so all these disparate pages would get collected by the sales rep and then sent to headquarters who would then send it all to their outsourced data entry partner,

[00:13:57] [SPEAKER_00]: who would look at each of those pages on one screen and type them all into another screen.

[00:14:01] [SPEAKER_00]: And as you might imagine, that was fraught with errors.

[00:14:05] [SPEAKER_00]: It was time consuming and might take days for the round trip.

[00:14:09] [SPEAKER_00]: And then errors would have to be corrected out in the field by the sales rep, etc.

[00:14:13] [SPEAKER_00]: They challenged us, DJO was the company, challenged us to transcribe or pull all the information from those pages right in the moment, just as the pages were scanned in.

[00:14:23] [SPEAKER_00]: And we've been doing that for DJO now for five years without any downtime and turning validated, accurate information back around to the sales rep in within a few minutes, sending the pages in.

[00:14:38] [SPEAKER_00]: And so the sales rep can just wait for the result.

[00:14:40] [SPEAKER_00]: And if there was a missing signature or missing page or whatever it was, I can get the error corrected right in the field.

[00:14:47] [SPEAKER_00]: And okay, so that was how we developed the technology.

[00:14:51] [SPEAKER_00]: We met with another, they're called durable medical equipment companies, and this one's called Apria.

[00:14:57] [SPEAKER_00]: And they are a very big respiratory equipment company that will sell you a CPAP device, sell you things that you use in the home, but that you need a prescription for and that gets paid for by the insurance company.

[00:15:11] [SPEAKER_00]: Well, Apria has been a great customer of ours for the last three years, help us round out the technology.

[00:15:18] [SPEAKER_00]: They get, we now process about 28 million pages a year that comes over the fax, 28 million pages a year.

[00:15:26] [SPEAKER_00]: And up until we came along, each one of those pages was manually looked at so that the data elements that needed to go into the claim that got sent to the insurance company could be found.

[00:15:39] [SPEAKER_00]: So an order would come in, it could be six pages, could be 106 pages.

[00:15:44] [SPEAKER_00]: And a data entry person would look at them on one screen and, and then type the data elements, the fields into, into their order intake screen.

[00:15:54] [SPEAKER_00]: And so notable came along and we now are a technology that takes in all, all of those pages, finds all the data elements points to the ones that need to be corrected.

[00:16:06] [SPEAKER_00]: So that the data entry person just needs to look at the exceptions, doesn't need to go hunting for things.

[00:16:12] [SPEAKER_00]: There may be three exceptions, two exceptions or no exceptions, but make sure it's all correct.

[00:16:17] [SPEAKER_00]: And then presses send.

[00:16:19] [SPEAKER_00]: And it takes the data entry time from what it was to something that is 5% of what it was.

[00:16:26] [SPEAKER_00]: So it's, we realized when we put that system in place, so we have something that was ready to go to market.

[00:16:30] [SPEAKER_00]: And that's what we did last year in September, started building a sales team.

[00:16:35] [SPEAKER_00]: And now we have many customers who are using our product in all sorts of ways in the medical arena.

[00:16:42] [SPEAKER_01]: Wow.

[00:16:42] [SPEAKER_01]: 28 million pages.

[00:16:43] [SPEAKER_01]: That is absolutely phenomenal.

[00:16:45] [SPEAKER_01]: And obviously AI is the flavor of the hour at the moment.

[00:16:48] [SPEAKER_01]: Everyone's talking about it, trying to work out.

[00:16:50] [SPEAKER_01]: It's great, but what is it going to do for me?

[00:16:52] [SPEAKER_01]: What is it going to do for my business?

[00:16:54] [SPEAKER_01]: Can you talk me through the moment where you started applying AI to transform healthcare?

[00:17:00] [SPEAKER_01]: Healthcare, at what point did AI come into the frame here and how are you using AI now?

[00:17:07] [SPEAKER_00]: So AI, first of all, is not new.

[00:17:10] [SPEAKER_00]: In fact, way back, even before AOL, when I was in graduate school for fun, I wrote an AI program that played the game of Bridge.

[00:17:20] [SPEAKER_00]: And so AI techniques have been around for a long time.

[00:17:23] [SPEAKER_00]: And the recent attention has been drawn to this new form of AI called large language models in the form of ChatGBT or Gemini or Cloud, all these different ways of using large language models.

[00:17:38] [SPEAKER_00]: So LLMs.

[00:17:39] [SPEAKER_00]: And so when we started Notable back in 2017, we used all sorts of techniques.

[00:17:46] [SPEAKER_00]: We have machine learning built into the system, natural language processing built into the system.

[00:17:52] [SPEAKER_00]: We have our own techniques.

[00:17:54] [SPEAKER_00]: We have this technique called spiders that goes and finds a patient name wherever it is.

[00:17:59] [SPEAKER_00]: Give it a page, give it 100 pages, and it'll find the patient name, date of birth, address.

[00:18:05] [SPEAKER_00]: Our spiders do that.

[00:18:07] [SPEAKER_00]: And you can think of those as AI, artificial intelligence, because they act intelligently.

[00:18:12] [SPEAKER_00]: And the techniques involved are quite sophisticated.

[00:18:15] [SPEAKER_00]: And so large language models have now been baked into our platform as well.

[00:18:21] [SPEAKER_00]: And one of the things an LLM can do is take prose, because clinical notes, for instance, and summarize it.

[00:18:28] [SPEAKER_00]: What's going on here?

[00:18:30] [SPEAKER_00]: And the answer to your question, Neil, what can AI do for you?

[00:18:34] [SPEAKER_00]: One of the things that it can do for you now is it can give you great summaries of lots of information.

[00:18:41] [SPEAKER_00]: Now, it's not flawless.

[00:18:44] [SPEAKER_00]: You probably heard this.

[00:18:45] [SPEAKER_00]: You probably heard the expression hallucination.

[00:18:48] [SPEAKER_00]: So sometimes LLMs, one of the drawbacks of LLMs is they sound human sometimes.

[00:18:55] [SPEAKER_00]: So when they answer you, it sounds like a human.

[00:18:57] [SPEAKER_00]: And we have a natural tendency to believe humans.

[00:19:03] [SPEAKER_00]: And so something that sounds human, we're a little bit more prone than usual to accepting something that's wrong, something that hasn't been thoroughly.

[00:19:13] [SPEAKER_00]: So one of the things that we are notable is very, very big on.

[00:19:19] [SPEAKER_00]: And I do a fair amount of speaking about AI.

[00:19:23] [SPEAKER_00]: I'm a fellow at Harvard and a fellow at Northeastern University in their AI institutes.

[00:19:29] [SPEAKER_00]: And I'm very adamant that CEOs, for instance, or a board that asks the question that you just asked, how can AI help me?

[00:19:37] [SPEAKER_00]: What's the thing is when you're talking to a vendor, somebody's selling you AI technology, you need to ask them, how does your system know when it doesn't know?

[00:19:47] [SPEAKER_00]: Because no AI system is perfect.

[00:19:49] [SPEAKER_00]: And it's really a science in and of itself to determine for a system to determine its level of confidence in what it is saying.

[00:20:01] [SPEAKER_00]: So, and LLMs aren't very good at that.

[00:20:04] [SPEAKER_00]: LLMs will just tell you.

[00:20:06] [SPEAKER_00]: They won't say, now, I'm not sure about this, but it doesn't really.

[00:20:10] [SPEAKER_00]: And in my business, you know, I'm working for notable systems is working for folks in the healthcare industry where accuracy is very, very important.

[00:20:20] [SPEAKER_00]: And so we've developed techniques for when we extract a piece of information and we may have used an LLM to extract it.

[00:20:28] [SPEAKER_00]: We have some way of either double checking or we have different methods for measuring our level of confidence.

[00:20:38] [SPEAKER_00]: And if for much of the information that we extract from documents, we have a high level of confidence and we'll just green light it.

[00:20:45] [SPEAKER_00]: We'll have a green check mark, no need to check.

[00:20:48] [SPEAKER_00]: And then other things, we might have a yellow check mark that says, pay attention to this, it's shaky.

[00:20:53] [SPEAKER_00]: Or a red check mark that says, you know, this is either missing or was illegible or, you know, our system doesn't know.

[00:21:01] [SPEAKER_00]: And every AI system needs that.

[00:21:03] [SPEAKER_00]: Every AI system needs that kind of wrapper to double check itself so the human knows when to pay attention.

[00:21:11] [SPEAKER_00]: That goes for autonomous vehicles, by the way.

[00:21:14] [SPEAKER_00]: You know, an autonomous vehicle needs to say, I'm a little out of my depth here.

[00:21:20] [SPEAKER_00]: It's raining.

[00:21:21] [SPEAKER_00]: There's traffic.

[00:21:22] [SPEAKER_00]: There's geese walking across the street or whatever.

[00:21:25] [SPEAKER_00]: Please grab the wheel.

[00:21:27] [SPEAKER_00]: So all AI systems need a way to alert the human when it's time to pay attention.

[00:21:33] [SPEAKER_01]: Completely agree with you.

[00:21:35] [SPEAKER_01]: And the hallucination thing is so apparent.

[00:21:38] [SPEAKER_01]: Recently, I was doing some research for an article and I was referencing a company and I needed to know how many users that company had got.

[00:21:45] [SPEAKER_01]: And there was conflicting reports online.

[00:21:47] [SPEAKER_01]: Some were saying 10 million users.

[00:21:49] [SPEAKER_01]: Others were saying 20 million.

[00:21:51] [SPEAKER_01]: So I asked, I think it was Claude AI, I said, how many users does this platform have?

[00:21:56] [SPEAKER_01]: And it came back with me and said, hi, Neil, this platform has 50 million users.

[00:22:01] [SPEAKER_01]: So my next response was, can you give me a source for that stat?

[00:22:04] [SPEAKER_01]: Because I've not seen anything anywhere close to that.

[00:22:07] [SPEAKER_01]: And I've just pulled this up on screen here.

[00:22:09] [SPEAKER_01]: And the actual reply it gave me was, Neil, I apologize for the confusion in my previous response.

[00:22:15] [SPEAKER_01]: Upon reflection, I realized I don't really have a reliable source for the 50 million users statistic I mentioned.

[00:22:22] [SPEAKER_01]: It lied to me and I just almost admitted to it straight after.

[00:22:27] [SPEAKER_00]: And it's funny that you challenged it and then it actually confessed.

[00:22:32] [SPEAKER_00]: That was pretty, that's pretty good.

[00:22:34] [SPEAKER_00]: And a lot of companies, so we use Anthropic and, you know, a lot of the open AI, LLM companies are now putting work into the double checking and the framing of things in terms of how much certainty, how much confidence there is.

[00:22:51] [SPEAKER_00]: Because, of course, it was immediately noticed when LLMs came out of the gate, you know, a couple of years ago that they were fraught with this hallucination problem.

[00:23:02] [SPEAKER_00]: And so it's quite important.

[00:23:05] [SPEAKER_00]: And especially in the healthcare arena, super important for a system to alert the user, hey, check that.

[00:23:12] [SPEAKER_01]: And I think one of the great things about AI is it helps people work smarter, not harder.

[00:23:17] [SPEAKER_01]: And there's this whole argument at the moment about AI removing jobs and we're hearing about it more and more.

[00:23:23] [SPEAKER_01]: And I always say that the magic happens when you have AI and humans working alongside each other, when they're both doing things that the other cannot do as well.

[00:23:31] [SPEAKER_01]: But together, that's where the magic happens.

[00:23:33] [SPEAKER_01]: And one of the things that, again, that put you on my radar is that knowable systems, you're applying AI to this health data platform intended to transform healthcare IT, but by humanizing health data input and access as well as automate the document processing.

[00:23:51] [SPEAKER_01]: That humanizing aspect of it, is that something that was very important from the outset?

[00:23:55] [SPEAKER_00]: That's well said, Neil.

[00:23:57] [SPEAKER_00]: It is very important.

[00:23:59] [SPEAKER_00]: And it's very important for our customers to think of it as a collaboration between technology and their people.

[00:24:09] [SPEAKER_00]: And we do not advise our customers, especially in the healthcare arena, to let go of the reins.

[00:24:16] [SPEAKER_00]: Simply plug and play, set it and forget it.

[00:24:19] [SPEAKER_00]: That is not what should happen.

[00:24:21] [SPEAKER_00]: The data is very important.

[00:24:23] [SPEAKER_00]: Keeping it correct is of vital importance sometimes.

[00:24:27] [SPEAKER_00]: It's the difference between a patient receiving a needed device and not receiving it.

[00:24:33] [SPEAKER_00]: And so we have this acronym for one of our products called AI-assisted data entry and AIDN.

[00:24:43] [SPEAKER_00]: So this AI-assisted data entry, or AIDN for short, is exactly what it says.

[00:24:49] [SPEAKER_00]: It's AI-assisted data entry.

[00:24:52] [SPEAKER_00]: It's not automated data entry.

[00:24:54] [SPEAKER_00]: And we do not encourage our customers to think of it as something they can just plug in and then the people can go on to other things.

[00:25:02] [SPEAKER_00]: There still needs to be someone mining the data store.

[00:25:06] [SPEAKER_00]: And as you said, it then enables you to get other things done.

[00:25:10] [SPEAKER_00]: And it becomes a very powerful when you go from having to hunt through 40 pages to find 17 pieces of information to correcting one or two issues.

[00:25:20] [SPEAKER_00]: You have a lot of other time to spend on other things.

[00:25:24] [SPEAKER_00]: And the machine can take care of what humans aren't so good at, which is repetitive remedial tasks to allow the humans to do what machines aren't so good at.

[00:25:33] [SPEAKER_01]: As you mentioned, when we very first started our conversation today, Notable Systems this year has raised $8.8 million in new funding for your Series A round.

[00:25:44] [SPEAKER_01]: Can you tell me a bit more about that and ultimately what it means for Notable Systems?

[00:25:49] [SPEAKER_00]: So just about exactly a year ago today, we were a small shop with three anchor customers or sandbox customers that helped us develop the platform.

[00:26:01] [SPEAKER_00]: We started hiring a small number of people on the sales side and then a large number of machine learning specialists, PhDs, engineers, so that we could keep extending the platform.

[00:26:16] [SPEAKER_00]: And we went to market and saw an immediate appetite.

[00:26:20] [SPEAKER_00]: Every single medical equipment company we talked to wanted what we had because they all had the same problem processing a swamp of pages and it needed to be done manually, either in-house or abroad.

[00:26:33] [SPEAKER_00]: And as soon as they heard what our system could do, they really got in line.

[00:26:38] [SPEAKER_00]: And we started closing customers.

[00:26:40] [SPEAKER_00]: We went to our founding investors and also an outside venture capital firm called GrowTech, Washington, D.C. based.

[00:26:49] [SPEAKER_00]: And we raised the $8 plus million so that we could grow the team faster and build our development capacity and our onboarding capacity.

[00:26:59] [SPEAKER_00]: So we really wanted to be able to take all kinds of customers, not just customers that wanted our standard product, but also customers who could help us build our product line.

[00:27:12] [SPEAKER_00]: So we have a customer that is where we are developing a brain to go along with the data.

[00:27:20] [SPEAKER_00]: So historically, our bread and butter product at Notable has been extracting information.

[00:27:25] [SPEAKER_00]: What about using the information?

[00:27:27] [SPEAKER_00]: Okay, well, and there's all sorts of ways we can bring LLMs into the game of analyzing the information, try to figure out what's going on.

[00:27:37] [SPEAKER_00]: And so we're building what's called a payer greenlighting system, which when you've got a claim just about ready to send to the payer, let's say it's Medicare, you want to know whether it's going to be approved.

[00:27:50] [SPEAKER_00]: You'd kind of like to know that before you send it in.

[00:27:52] [SPEAKER_00]: So you fix whatever problem it was, or if it's not complete or it's not in compliance somehow.

[00:27:58] [SPEAKER_00]: And so what our payer greenlighting system does is it alerts you.

[00:28:02] [SPEAKER_00]: It actually does the analysis and it alerts you to alerts our client to the probability it'll be approved.

[00:28:12] [SPEAKER_00]: And then where the gaps might be, there's another page missing, or you need the letter of medical necessity or whatever, whatever the issue is.

[00:28:21] [SPEAKER_00]: It'll light it up and say, you know, finish, finish the order before you send it in.

[00:28:25] [SPEAKER_00]: And that kind of brain, that kind of analysis, we're developing in conjunction with a couple of our customers.

[00:28:32] [SPEAKER_00]: And so we have customers not only using our standard product, but essentially helping us as a sandbox to extend our product line.

[00:28:41] [SPEAKER_00]: And all of that, of course, requires investment.

[00:28:43] [SPEAKER_00]: And so we raise the money in order to build out the team, the innovation team.

[00:28:48] [SPEAKER_00]: I wouldn't be surprised if early next year we're raising more money because we're seeing that demand is extraordinary.

[00:28:56] [SPEAKER_00]: And so there's a huge growth potential here.

[00:28:59] [SPEAKER_00]: We're really excited about it.

[00:29:01] [SPEAKER_01]: Wow, exciting times ahead.

[00:29:03] [SPEAKER_01]: And just to bring to life everything that you're talking about here, do you have any stats, use cases, or anything that will allow anybody that's in this industry,

[00:29:12] [SPEAKER_01]: maybe a future customer, the kind of things that they can expect to save, the problems that they solve, et cetera,

[00:29:20] [SPEAKER_01]: the difference that it will make to them and their business or healthcare department.

[00:29:25] [SPEAKER_01]: Is there anything that you can share that would just help them understand where they are now and where you might be able to take them?

[00:29:32] [SPEAKER_00]: Well, you know, I ask every customer, every prospect we talk to, what do they think the value is?

[00:29:41] [SPEAKER_00]: And my favorite answer actually came from a company that's now a client called Aeroflow, who said patient satisfaction.

[00:29:50] [SPEAKER_00]: Patient satisfaction is their mantra and notable will help bring their patients more satisfaction.

[00:29:57] [SPEAKER_00]: And how are we going to do that?

[00:30:00] [SPEAKER_00]: Well, by helping them turn their order around faster.

[00:30:05] [SPEAKER_00]: So order comes in, patient gets referred, Aeroflow, because of our technology, can complete the claim, get the insurance reimbursement faster, approve the order, ship the product much faster.

[00:30:19] [SPEAKER_00]: And reduce denials.

[00:30:21] [SPEAKER_00]: So this brain system that I'm talking about, our payer greenlighting, also helps make sure you don't have a false start with the insurance company.

[00:30:30] [SPEAKER_00]: You don't want to send it in and then wait a week and then it gets denied or it gets provisionally approved because something was missing.

[00:30:37] [SPEAKER_00]: And so all of that goes to patient service, patient satisfaction.

[00:30:41] [SPEAKER_00]: Now, of course, the economics of delivering that improve also because it takes fewer people to process the claim.

[00:30:50] [SPEAKER_00]: You do it faster with fewer mistakes.

[00:30:53] [SPEAKER_00]: And so there are all sorts of elements economically that make it cheaper and faster and more efficient for the medical equipment company to process the information.

[00:31:03] [SPEAKER_00]: So all of that goes into our value proposition.

[00:31:06] [SPEAKER_00]: And that's why there's such an appetite for a product among the DMEs, the durable medical equipment companies, because it gets them all of that.

[00:31:14] [SPEAKER_00]: Happier patients.

[00:31:15] [SPEAKER_00]: More referrals because the folks referring patients to them start to get the word.

[00:31:21] [SPEAKER_00]: You know, which companies are the ones that have the fastest turnaround and the best approval race?

[00:31:27] [SPEAKER_00]: And so it helps them as it helps them competitively as well.

[00:31:31] [SPEAKER_01]: I think that's a powerful moment to end on.

[00:31:34] [SPEAKER_01]: I love how this conversation has all been about technology, but the end result is all about happier patients.

[00:31:41] [SPEAKER_01]: And for anyone listening wanting to find out more about anything we discussed today or equally connect with you or your team, if they've got any additional questions, where would you like to point anyone listening?

[00:31:52] [SPEAKER_00]: Please come to our website.

[00:31:54] [SPEAKER_00]: It's NotableSystems.com.

[00:31:58] [SPEAKER_00]: And plenty of demos there, plenty of information there, and then a way to contact us by email if you would like to.

[00:32:07] [SPEAKER_01]: Awesome.

[00:32:07] [SPEAKER_01]: Well, I'll add links to everything so people can get in touch with you nice and easily.

[00:32:12] [SPEAKER_01]: And as I said, it's very important to say talking about the technology, but the real world impact that it's having on liberating human potential through automated document processing and ultimately leading to happier patients.

[00:32:25] [SPEAKER_01]: And just a big thank you for shining a light on this.

[00:32:28] [SPEAKER_01]: We hear too much about the hype sometimes, but not about the real difference it's making.

[00:32:32] [SPEAKER_01]: So thank you for sharing your story today.

[00:32:35] [SPEAKER_00]: Thank you very much, Neil.

[00:32:36] [SPEAKER_01]: As we close our conversation with Steve today, I think it's clear that Notable Systems is doing more than just automating healthcare processes.

[00:32:45] [SPEAKER_01]: They're actually transforming the very fabric of healthcare IT.

[00:32:49] [SPEAKER_01]: And doing that by combining advanced AI with essential human oversight and maintaining that human oversight is not just improving efficiency, but also ensuring that patient care remains at the heart of the innovation.

[00:33:03] [SPEAKER_01]: I think that is the important part of it.

[00:33:05] [SPEAKER_01]: And their impact of their work is already being felt with faster order processing, fewer insurance claim denials, and a significant reduction in manual tasks that often bog down healthcare professionals.

[00:33:18] [SPEAKER_01]: And as I find myself reflecting on today's discussion, I think it's worth considering how AI can be thoughtfully integrated into our systems to truly enhance, not replace human effort.

[00:33:32] [SPEAKER_01]: That is the big takeaway for me.

[00:33:33] [SPEAKER_01]: And as they continue to grow, supported by that Series A funding, I think their journey offers valuable insights into the future of healthcare technology.

[00:33:43] [SPEAKER_01]: How will these advancements shape the experiences of patients and providers alike?

[00:33:49] [SPEAKER_01]: Something for us all to ponder on.

[00:33:51] [SPEAKER_01]: Remember, you can connect with me on LinkedIn, Twitter, Instagram, just at Neil C. Hugh.

[00:33:55] [SPEAKER_01]: Send me a DM, let me know your thoughts, and have a think about the potential of AI in your own industry.

[00:34:01] [SPEAKER_01]: Consider the balance of innovation and the human touch that makes it all possible.

[00:34:06] [SPEAKER_01]: Other than that, I'll be back in your podcast feed tomorrow with another guest.

[00:34:11] [SPEAKER_01]: So thank you for listening as always, and hopefully we'll get a chance to speak again tomorrow morning.