Oh the Humanity! (in the loop)

Mindy

Corbett: Hi everyone, we are back. It's Kay and Mindy back with another episode of Now That I Know. This week we're going to be talking about AI and the human gatekeeper. basically talking about how we can use AI in a way that's effective, but also making sure that we are compliant and we have controls in place. And we're making sure that there's always a human in the loop. with that in mind, I am going to kick it over to Kale and see what he's got to add. And here we go. All you

Cale: No, thank you. Hey, all right. So like Mindy mentioned, you know, we're gonna talk about what what I might call governance and and human in the loop and all this kind of thing related to AI. I'm gonna be teeing her up some questions specifically related to denials and and revenue cycle management and things like that. And she might punt it back over to me at some point about, you know, a different take on on AI and and data and that sort of thing. I think the one thing that I want to kind of kick off with, Mindy, when we talk about the human in the loop aspect. it's not just to me the human in the loop, it's the human in the beginning, the human in the loop, the human in the process, and the human in the end, you know, with the output. And so to me, I think as we're talking through this, if if it's true for you in the RCM space like it is for me elsewhere. We need to Mindy

Corbett: Mm-hmm.

Cale: really remember the human has to be involved from literally start to finish. We're not really best served by turning over our businesses, our data, our our our well-being over to machines at any point. And we're we've seen you and I both seen numbers coming from the corporate world where they're finding out the ROI just isn't there. they're finding, of course, data governance issues and and on and on. And you live in a space where not only is it is it the ROI aspect, but it's also HIPAA and a lot of the other regulatory governance situations that are very critical and very integral to how the AI pieces itself is governed in regard to that data. So in that roi space. You know, there's been the expectation that AI was going to help us reduce costs, speed things up, you know, easier, better, faster, cheaper, right? So in Mindy

Corbett: Mm-hmm.

Cale: that context, the first question I had, what are the expected reductions in the number of denials, the amount of rework, and then like you I think you call it revenue leakage in that, you know, the healthcare RCM revenue cycle management. how do how do you measure success in your world with AI in the mix? Where where is it coming in? What's its real impact? You know, talk to me about what does success look like so far. Mindy

Corbett: Mm-hmm. Well, so it depends on how you treat it. It's it's really it's a matter of if you're using it in a way that's smart and you're using it to automate tasks that can be, you know, you automate repetitive tasks and you keep a human there in the loop on things that require somebody to look at for compliance person purposes then you can actually get a lot of really good return on investment with them. But you have to really, really make sure that you're careful with that. So you can see a great decrease in denials if you're using them on the if you're using AI on the front end, for example. So one of the things that I do is I have a tool that looks at claims and says, Based on historical data and based on predictive analytics and based on the way that this claim this particular claim going out the door is set up, is there a high probability that something on here is going to be denied? That's a great way to use AI, right? That can say, right? So it's great for things like predictive analytics. It's great for making suggestions. It's great for that pre-claim scrubbing. So I use it for, like I said, that pre-flight claim it claims engine that I have. it's great for things like intelligence to be able to tell you, you know, this this was denied, this is why. Let's push these into categories and make them intelligent categories, make them categories that actually make sense. it's wonderful for stuff like that. So if you can take those things that are repetitive. that can do some predictions and and use it for things like that, you can see a lot of return on on investment in that. I mean, you can see up to seventy percent return on investment for things like that.

Cale: No. Mindy

Corbett: You have to use them in a smart way.

Cale: Let me ask you, no, that makes perfect sense. And and the question I'm about to ask you may not have a may not have a basis in reality in terms of something that can be done, but follow along with me here, hopefully. y you were talking about in the way I would think of this is, you know, you're talking about at the beginning of the of the revenue cycle. You're talking about, you know, bringing AI to bear as early as possible to to manage and govern that data and get it through the right channels as quickly as possible with human guardrails and governance and all that. My Mindy

Corbett: Right.

Cale: my my question and and that and there's already a cost incurred to some degree to that point, right? We've got denials in the system already at that point. And so there's some Mindy

Corbett: Mm-hmm.

Cale: cost that's been that's been borne and we're trying to reduce the cost from there forward. Now where my brain is going next is from the situation of What what is the industry doing to prevent the denials in the first place? Is there something that can be done that leverages AI that says let's be smart as far back as possible so that the denial never happens? You know, talk to me about maybe yeah. Mindy

Corbett: That's kind of what yeah. That's kind of what I'm talking about. So the that that one one of the tools that I've built, so it's it's an automated module that I built, will actually do that. So it takes it takes information from a couple of different places. So one, it looks at historical denials. Two, it looks at common issues. that get so common reasons for denials. So modifier anomalies, common, common procedure codes that get denial, get to get denied for different reasons. It looks at things like provider enrollment. and is this gonna flag a denial? And so what it's doing is this module will actually look at claims before they go out the door. So it takes your standard claims scrubber and it basically puts it on steroids. And says, take everything a standard claim scrubber does and do more. And tell me of all of these claims that are going out the door, which ones of these have a high potential to be denied. Not just that have some errors that I need to fix, which ones have a high potential to be denied so that I can figure out why and I can fix them before they go out the door so that I prevent. As many of those denials as possible before they go out. So that is a really

Cale: Okay. Mindy

Corbett: great use of AI because it can

Cale: Okay. Mindy

Corbett: be effective that way. And you can actually

Cale: That's okay. Mindy

Corbett: and where the human comes in there is the human then on the back end says, I've been seeing a lot of this particular issue get denied. Let's feed this back in the front end so that it catches it on the front end next time.

Cale: There you go. Okay. I was just gonna ask you the the whole point of this is to is to catch it early and then Mindy

Corbett: Mm-hmm.

Cale: make it smarter so that when I now wanna I don't want to oversell this, but it's not only making the system smarter, but if I can, you you talked with me about this offline a couple of weeks ago. Your other Mindy

Corbett: Mm-hmm.

Cale: goal is to make the human smarter so that he or she They know intuitively now by experience, they know by seeing this data, they know sooner than later when they're keying or coding in, they're capturing that that that billing situation from the provider, they know Mindy

Corbett: Mm-hmm.

Cale: as early as possible how to do that more smartly than they did previously. Mindy

Corbett: Mm-hmm. Mm-hmm.

Cale: So there's this this but feedback loop, I guess, if of sorts, you're making both smarter, the system and the Mindy

Corbett: Yeah, yes.

Cale: person. using it right. yeah, Mindy

Corbett: Right. Right.

Cale: so I would be very interested in and we'll put we'll put links and please do make sure we put links in the show notes and this type of thing. I want people to be able to s literally see what you're talking about that early on because I you know what we've been hearing you know in the press and we're seeing in our own you know practices if you will is we're seeing a lot of these horror stories coming and we're hearing The promise of AI is going to make things better further back. Mindy

Corbett: Mm-hmm.

Cale: there aren't as many people in my mind showing us that happen. So I get excited anytime I hear, you know, someone like you saying, Yeah, we are thinking as far back as we can and making everybody smarter. that I like that. yeah. Mindy

Corbett: Right. That's how we need to use it. We need to, we need to be smart about it. There are a lot of situations where you have to be very careful, but use it for things that are repetitive. There's absolutely no reason why, you know, why I have to sit there and manually do things that a machine can do if there is somebody on the back end who can validate that all of that was done successfully and in a compliant manner.

Cale: Well, I'm I'm glad you mentioned that because one of the things that I wanted to kind of have you tell me more about and and help me understand from a from a help me be comfortable. I'm s I'm ignorant of the details of RCM and healthcare, right? So when I when I ask you this question, I'm asking on behalf of the people I've seen as your customers, helping. You know, we go out and we see in the message boards, the the LinkedIn pages, you name it, we see what I would call a lot of doom and gloom from the men and women on the ground doing the coding and billing, for example, talking about how AI either is going to or has replaced them, or there's certainly a fear that that's going to happen. So let me kind of have you talk me through. From the industry perspective, as b as best you can tell with your with your crystal ball on, and then also too, from the perspective of a person with a solution for RCM that uses AI, help me understand what it's going to look like for different players, different humans in this Mindy

Corbett: Mm-hmm.

Cale: RCM loop that are you know, in this R CM flow that you're in, the revenue cycle. Help me understand where the humans are going to continue to be useful and be needed versus where we can give it to the machine and and why that's good. Mindy

Corbett: Yeah.

Cale: You know, talk to me, get me off the doom and gloom bus, right? And get me on to the bus of, you know, what's gonna happen? Mindy

Corbett: Yeah, so I think there I think there is a lot of fear because it's a it tends to be a fear-based role anyway, because everybody feels like they're replaceable. And anytime you hear about, you know, something coming in and taking over, it sounds really bad. But the my feeling is, and I found this in my own life doing RCM, anytime you can bring in an automated process. That means that you're freed up to do higher level tasks. And so what then that means ultimately is you're freed up to do something that generates revenue. Right. So instead of being stuck hunting and pecking and looking for little bits of money here and there or chasing the same denial over and over again, you're you're off doing things that actually maybe are bringing in money, looking at that higher dollar stuff. Instead of just little penny annie stuff. So what's happening is instead of being the person who's sitting there looking at every single individual denial, you have something that's automatically categorizing all of your denials for you so that you are focusing on the ones that are recoverable, that are high value, so that the ones that you are fighting are the ones that are worth money instead of the $5 ones.

Cale: Okay. Okay. Yeah. Mindy

Corbett: Right. Right. And then the other thing that that's happening is when claims go out the door, if you've got something that's automating your eligibility checks and it's doing it well, instead of doing what some of the clearinghouses are doing now, where it's it's half-assed and it's not always successful and it's not always 100% accurate. You know, you've got something that's doing your eligibility checks and it's doing it well. You've got claim scrubbers that are checking things before you go out the door that are doing a hundred percent better than what's going on now. So you know when claims go out, they're going out clean, they're going out accurate, they're going out with a lower likelihood of being denied. Money's coming through the door better. So that the stuff that you're going after on the back end is higher quality.

Cale: we talked about this a little bit at the beginning, and talked about it offline as well. And that's the concept of data governance, right? data protection, data cleanliness, you know, all the stuff that says, look, garbage in, garbage out. We don't want bad actors getting in the middle of it. We don't want it being stolen, you know, HIPAA, blah, blah, blah. is your concern about AI in terms of what it could do to the data out there for your customers, your providers, the patients like yourself and me? What keeps you up at night about the data in AI? But then maybe just as importantly, in your own practice there where you've got ROI that works. What are some of the steps that you've taken with the with the the product that do protect the data, that govern it and and so on? Mindy

Corbett: I mean, my biggest concern with anything, anything where there's data going back and forth when you're talking about healthcare data, PHI, is a breach. You know, you're always concerned about that. And with AI, you don't always know what's gonna happen, but there are regulations in place to make sure that there's a clear audit trail. You have to have a named person accountable for that. So you have to have a HIPAA compliance officer, period. you have to have somebody who's assigned, who's who's responsible for the governance of that AI data or anything that goes in and out through that AI. And then for my own personal business, my platform, anything that is being done directly on the website, there is no PHI exchanged, period. Don't there's no PHI

Cale: Okay. Mindy

Corbett: there at all. If you're using one of the automation modules, that is something that you will license directly. And for that, there is a BAA required. There's there's there's HIPAA documentation required. We go through HIPAA training. hi have a HIPAA compliance officer, all of that stuff is buttoned down. So it's all on the website under our security information. If anybody ever needs to read that, it's all available there. but that is all covered under the the BAA.

Cale: This is something I think about all the time, and it's not just in in RCM or any other space, it's data protection, period. You know, you talked about having, you know, a named person responsible and all the different regulations and and that kind of thing, which are great, except to my mind that's that's after the horse may have left the barn. So sure, you've got a person who's got a name they're responsible for when crap goes bad, data gets stolen, breached, blah, blah, blah. You know, Mindy

Corbett: Uh-huh.

Cale: fine. So that's the person who's going to get in trouble, if you will. that's that's fine and dandy, but what's the situation, you know, and again, I'm thinking maybe broader industry and then, you know, at your level, what are y'all doing to keep the bad guys from even getting to the data in the first place? Mindy

Corbett: there needs to be security measures in place. You know, you know, SOC two or whatever, you know, whatever your particular industry requires, there needs to be security measures in place. and and that's just that's just standard. and from our perspective, you know, from a from a compliance perspective and healthcare, and as far as the AI Keeping that compliant, that person, that names person is responsible for monitoring that and validating that it does not get out.

Cale: Now and so that's on a there's a there's a a cadence to how they keep they and their team keep track of of the data and all that. Yeah. That makes that makes sense. I think the I think the last question that I've got to kind of bring this all home for those of us who don't live in your world, you know, to that level of detail is where do you want to see the the revenue cycle management industry go. So let's say five years from now, after the buzz and the hype and all that has settled down and AI is just part of the fabric, let's say, Mindy

Corbett: I've got it.

Cale: where do you want to see the industry having gotten to through all that? what's a next stage look like for you and your industry when this is all done? Mindy

Corbett: I mean, I'd like a level playing field. I'd like it to be a situation where payers aren't using it against practices to to deny things before they actually ever get a chance to see the claim. I would like it to be a situation where both sides are equal, where it's not this, you know, we know this provider, we don't care what they send us, we're gonna deny it because we have AI that says we can. I I right. I I don't want

Cale: Yeah. Yeah. Mindy

Corbett: any more of that. I want that playing field to be level. I would like it to be a situation where we're not just using AI for AI's sake. I want it to be used in a way that makes lives easier. So use it where you can automate things that are repetitive. You can use it pr to predict outcomes. You can use it for forecasting. You can use it for workflows and frameworks and optimization. And then you have your team who is still there who can come in and say, This doesn't look right. Let's validate that the AI isn't hallucinating.

Cale: Right. Right. Mindy

Corbett: And you still that have that hu human who's still making that final decision because AI is not always right. It's not perfect. You still have humans for exceptions. You still have inj humans there for a compliance because that's always important. And that human is always help that AI keep learning and keep improving. So in in my head, it's it's there's always gonna be a balance between the two. And so in a perfect world, the AI helps us do better. We help the AI do better and the playing field is level between us and the payers.

Cale: That that makes I I like the sound of that world and there's something you said there toward the beginning of that of that you know, laying that out that that I wanted to circle back to I hate to take you back past the part where we were using that as a conclusion, but it you said something that I want to that I want to dig into a little bit. I saw an article a couple of different times and I when I say a couple weeks ago, it might have been a month or more ago. And I brought it to your desk and it was something to the effect of, and I was asking you to validate this for me because again, I don't live there, but it was something to the effect that the payers, and we say payers, I think the English translation of that is insurance companies and so forth, who Mindy

Corbett: Yeah. Yeah.

Cale: pay the pro pay the provider, meaning the doctor, right, for your health care, okay, that they deliver to you. Right. I saw something that was talking about the payers, the insurance companies, using AI to, for lack of a better term, predictively ahead of time, if there is such a thing, deny claims, go out of their way to use AI in what I might call a weaponized fashion, in order to make it more difficult or slower for. Mindy

Corbett: Mm-hmm. Mm-hmm.

Cale: your physicians, your providers, men and women to get paid for the services they've provided, you know, net forty-five, net sixty, net Mindy

Corbett: Yeah.

Cale: whatever, right? Now, what is the truth? How much of that is height versus how much of that is truth as far as you know so far between this weaponized AI versus what's really happening on the ground? Any any read so far? Mindy

Corbett: know the exact number, but I do know that they are using AI to slow down payments. so they're What's happening is, and this is a real thing, they are denying claims in an attempt to pay slower. And what is happening is something in the order of 80%, 90% of those claims are ultimately getting paid.

Cale: Okay. Mindy

Corbett: if they deny a hundred of them only about ten of them should actually have been denied is what's happening and so yeah they are denying claims in an attempt to be able to pay them slower

Cale: know, I I know you and I, we may be neck deep in our respective domains and keeping track of of hype versus reality or fear versus reality or so forth. So when I hear things like that, one of the very first things I ask myself is. How much of that is true? how much of it's not, and you know, who who benefits from putting that that out there, you know, and yeah. Okay. Mindy

Corbett: I would just say the only thing I would say is don't be scared of AI. Use it. but don't abuse it. use it to your advantage, make it work for you, don't let it work against you. and if if you're afraid of it taking over coding in reality, it really can't. There's too many regulations around that for it to take over completely. It can attempt to make suggestions, but in the end, a human actually has to ultimately decide. in order to stay compliant. So don't let that fear overrun your life. So that that's kind of where I would go. I said, don't be afraid of it yet. let it work for you. Keep an eye on it, but at this point, let it work for you. Let it

Cale: Yeah. Right. Yeah. And that's part of being Mindy

Corbett: as a tool, but don't let it own you.

Cale: Well, we've got a friend we've got a friend named Chris who calls it spackle, right? You know, I might call Mindy

Corbett: Yeah.

Cale: AI a hammer. His his phrase is he calls it spackle. But the idea is that's the human who has to apply the Mindy

Corbett: Mm-hmm. Mm-hmm.

Cale: AI. And I think that's whether you it's a hammer, a a hacksaw, spackle, whatever tool you wanna call it, the the human is the key there, right? And that's and I think that's the thing we want to keep first and foremost. If you're if you're okay with it and you'll allow it, I'm gonna I'm gonna kind of go quote off script a little bit because there's something I wanted the men and women who who listened to this and watched this, so I want those seventeen people to hear hopefully compelling answers from you about the questions I'm getting ready to ask. So Mindy

Corbett: Okay.

Cale: take yourself, take yourself back to say December, January, you know, early this year, late last year. You were Mindy

Corbett: Mm-hmm.

Cale: you're sitting in your home office, you know, behind me, and Mindy

Corbett: Mm-hmm.

Cale: I came in to the to your office out of mine, and I came in to ask you something, and you said, I want to show you something. Now and I'm and I'm s one of the things that folks who know me know I speak in what I call themes. And so Mindy is going to resist correcting me on the details and she's gonna ride the theme with me. So I walk into her office and I swear it's something like this. She goes, I'm so glad you're here. She was very excited. She goes, I'm so glad you're here. And I'm like, well, thank you. She goes, Well, yeah, I want to show you some scripting that I've done on my GitHub. And I heard the word scripting and I heard the word GitHub and I stood there and I'm like I had never thought I was ever gonna hear the words scripting or GitHub out of my wife's mouth ever in my life. and what was so Mindy

Corbett: Yeah,

Cale: funny about that was it was a complete not only were the words coming out of her mouth a complete shock, but the the fact that she had been thinking and working on what became ROI that works in all the underlying logic and things for weeks or or so prior. Now, I do want to take one step back. I don't want to minimize your technical prowess and your experience. You've been building and doing down in the technical weeds the stuff that you do today that has now taken shape as your own business. You've been doing that literally both with your previous partner, you know, and then just here and there for five, six, seven years. This you're not new to the technical layer and the problem solving layer. I don't want people to misunderstand me. But you hadn't really talked with me about it. And I think that's one of the things that is we cover in this podcast is we sit so close, we have this overlapping knowledge base and we do a lot of the same things, but we don't really ever talk about them to one another. Right. And so Mindy

Corbett: Mm-hmm.

Cale: that was the part that really surprised it wasn't so much the fact that you knew from GitHub and knew from scripting, it was the fact that you were doing it, working on it, that just blew my mind. Right. And Mindy

Corbett: Yeah. Yeah.

Cale: I'd laid that foundation there just for people to see a couple of things. I wanted them to see that, hey, there's a technical, smart, plugged-in human woman on the on the back end of this thing. And the reason I say that is because that human in the loop, this woman is and was the loop, right? So I'm gonna with that with that foundation laid, I'd like for you, if you can, kind of talk with everybody about okay, you had this idea. You had this problem to solve. And you, yeah, admittedly, you used AI to help you put form to it. And I'll leave it there because I can just keep going and I can say your whole story for you. But I want you to sort of share your relationship to AI, your idea, your DNA compared to where AI plays in, because all too often we hear vibe code or, you know, the Mindy

Corbett: Yeah.

Cale: AI did it. Give me the give Mindy

Corbett: Mm-hmm.

Cale: us the reality on the ground of what Mindy's lived experience was to get her to this point today. Mindy

Corbett: Okay, sure. So the short version is everything that is live in my website is from my brain. Everything came from inside of my head. So I pull some things from source data, I translate it into human human language, and what AI did for me was format it. So if I didn't know how to make it look that color or be in that box or that shape, that's what AI did. But all of the data, all of the details, all of the information, all of the formulas, all of the categorization, all of the details, all of that was from my brain. So I used

Cale: Yeah. Mindy

Corbett: AI to help me build the scaffolding. But all of the different pieces of information, all of the segments, all of the all of the data, that's mine.

Cale: And and that goes down to, and don't let me mischaracterize, but when you were talking to me there early in the year, late last year, and you mentioned the scripting and the scripts and all that kind of stuff, that is the logic and the the the math or what have you that provides underneath those different solutions that ROI that works exposes to to its customers, right? Okay. Which I you know. Mindy

Corbett: Right. Right. Right. Right. So for example, if you are somebody who works in the RCM space and you understand what goes into an eight thirty five in ERA. If you understand basically what goes into an eight thirty five, not everybody knows all of the little technical pieces of it. I know every single segment and every single element within those segments and every single separator. And so I know how to speak to whatever tool I'm using and say, these are the different pieces of information that I need from that 835. This lives in this segment, in this element of this segment, and it is divided by this element separator. And I

Cale: Yeah. Mindy

Corbett: only need these pieces. Or so that when I say all of that data came from my brain, each one of those pieces of EDI information. or ANSI information or whatever, I know every single one of those segments where they live, where to pull them from, and how to extract that data. So if I'm working with AI, the only thing that they're doing is formatting it into an HTML file.

Cale: Well, let's la I want to pressure test you on that a little bit. not to not Mindy

Corbett: And even that I

Cale: yeah, not to take away I I am not gonna minimize the Mindy

Corbett: Mm-hmm.

Cale: the glory, anything of that nature. What I wanna what I wanna make sure I understand though is let's let's pretend for a moment, and I know enough about this from listening to you that these kinds of things have to be maintained and updated. The data underneath the formulas, all the all the the the databases and stuff are are refreshed. you know, Mindy

Corbett: Uh-huh.

Cale: from the system somewhere in the cloud on like a monthly basis. Your logic your logic is not sitting waiting for Mindy to refresh it by hand coming out of her brain. If you got hit by the metaphorical bus tomorrow, the the data would stay updated. I I as I understand Mindy

Corbett: Right. So most of the data that we're pulling for for healthcare organizations or you know in RCM is standardized data. It's all coming from standardized data sets. And so all of that data has a specific update cadence every, you know, ever either yearly or three times a year or four times a year, depending on what the data is. And so on

Cale: Or really whatever you have. Mindy

Corbett: that standardized data gets refreshed. And so the only thing Mindy is needed for is if there's any type of of human element that needs to be tweaked to make it human readable. but the standardized data is refreshed on whatever that cadence is from that standardized location.

Cale: Okay, okay, that that makes sense. because there's that deal where people are comforted, I think, by knowing that a person's out there not just turning AI into a business. There's you know, there's some thought and some and some experience behind it. And then the flip side is that danger of if it's resting on the shoulders of a small number of people, the whole thing could come crashing down and something that they're relying on for their business is so fragile because it's single Mindy

Corbett: Mm-hmm.

Cale: threaded. I didn't want anyone to come away with the impression that even though it's your brain's baby, that if you stepped away tomorrow, their hard earned Mindy

Corbett: No.

Cale: money and investment would suddenly be at risk. You know, that's the that's the takeaway Mindy

Corbett: Right. Right.

Cale: I didn't want people to to make, you know. So Mindy

Corbett: Yeah, no, I have I have basically for lack of a better word, I have feelers that go out and go, Okay, it's time for new cart codes. Grab those. let's refresh. Has anything changed? Has anything been deleted? Has anything been updated? Are there any is there anything new? Same thing. Payer policies. Check the major payers, check Medicare, Medicaid. Have there been any pa payer policy updates? Let's see if we've got anything we need to check. Then I of course I've got the the tool you the thing you built for me that does that signal check to see if there's anything out there that is important in the industry that is causing a ruckus. Let's

Cale: Right. Mindy

Corbett: let's see if we can build some education about that. so

Cale: Right. Mindy

Corbett: no there are that's where another way AI comes into play is going out there and grabbing new information so that we can build education

Cale: we won't deep dive on this, but I wanted I wanted to jump in there before before I lost the thought. y you've got annual lack for a lack of better term, annual regulations that come online, you know, the twenty twenty seven rules versus the twenty twenty six rules. And I I ran across a thing that signal thing you were talking about where I was starting to see signals coming from the people on the ground. They were seeing what they were interpreting as changes coming from the from next year's rules that people are getting a preview of, they thought they were being held to them now, when in reality they aren't and they can't, but there was this loop of misinformation Mindy

Corbett: Mm-hmm.

Cale: that again, your tool in the library and the in the knowledge base is is populated to sort of Mindy

Corbett: Mm-hmm.

Cale: educate them properly about, right? You're hoping to fight misinformation on let's say Mindy

Corbett: Yeah.

Cale: Reddit or LinkedIn or Facebook or TikTok or, you know, hippa dot com, whatever you know, whatever, right? Yeah. Yeah. Right. So I mean that that's the idea. Mindy

Corbett: Yeah, yeah. No. Yeah. And there are situations where right. There are situations where a payer will put out a policy and say, This is our policy. This is how we're going to process this claim. And we are enacting this rule, and that is their rule. And that is how they do it. And you have to understand that rule. But they you can't, you have to understand what that is. And if they're if they're Processing something contrary to that rule. That's winnable.

Cale: Okay, fair enough. All right. thank you for playing along with me and letting me kind of drag you through the history and some of the deeper, you know, reaches of how you got to where you were. Because as your as your hype man and your and you know, your hero, I'm trying everything I can to to help people understand who they're dealing with, you know. But yeah. So w what are you d huh? if you had anything that you wanted to flip around before we closed, with me in terms of human loops, AI, you know, you it floor's yours. Mindy

Corbett: U AI.

Cale: what's that now? Mindy

Corbett: Talk about you in AI.

Cale: Well, you gotta be a little more specific than that, you know. Mindy

Corbett: I mean, so we talked about how it applies in revenue cycle, how we're using it in healthcare. How do you use it in your business? what does it mean from a government's perspective for you? How do you keep a human in the loop or do you have to keep a human in the loop?

Cale: the short answer is we just like you, we should. and the smart players do. the ones who are destined to learn the hard way don't. And and now they're finding out. in that operations and governance layer that I play in for, you know, program management. in you know the tech st sector and the tech adjacent sectors and that kind of thing. there was a push, a really big push for let's you know, 20 late 2024 onward, we're gonna flip the switch. It's not going to be a gradual, smart, considered rollout of AI against our current structure. It was we're gonna burn the whole damn thing down. And we're going to replace it with AI, and we're going to whack we're going to whack the jobs of 12, 15, Mindy

Corbett: Mm-hmm.

Cale: 14, 20,000 human beings at a time, send them out Mindy

Corbett: Mm-hmm.

Cale: with a severance package and replace them with AI. Well, that's a great, that's a great dream. And I say dream because the reality a year and a half or so later, in some cases, for those folks, is not. Mindy

Corbett: Year and a half or so later in some cases for those folks is not two years and not three.

Cale: Two years is not really landing the way they thought. Mindy

Corbett: Yeah.

Cale: we're we're finding you you've read the same thing I have where they're like, the ROI is not there yet because two reasons. One is hey, they laid AI against their broken governance, their strategy that was not well thought out in the first place. So, right, you've Mindy

Corbett: Mm.

Cale: heard me rail and talk and and post and and write about, hey, If you've got broken governance or you've got a a crappy strategy, guess what? Mindy

Corbett: Right.

Cale: you now have a crappy strategy that you've given over to a machine. And if you don't have someone watching what that machine is doing, it's going to go off on a tangent at at theoretically with your crappy strategy and go places you had no idea. Right. it's going to put out information that's going to damage your company's reputation. It could run you headlong into situations with regulatory bodies, auditors, Mindy

Corbett: Mm-hmm.

Cale: The market, the competition, you name it. Mindy

Corbett: Mm.

Cale: just or data leaks. Who knows? The the permutations of pain could be endless. Was that? Mindy

Corbett: Weasels.

Cale: Weasels too. Weasels, ferrets, any of that, right? You're Mindy

Corbett: Yeah, right.

Cale: you could be in big trouble. There's a there's a story behind that that we'll talk about another time. She doesn't just say the word weasel loosely. and so that's one, right? And the one as a human being here that sort of that one scares me because that's one of those ones I, as a living in that governance and and operations and strategy space for as long as I have that one where crappy strategy, misshapen operations, so forth. That's bugged me and plagued me for a long, long time. These were all solvable problems before AI came along. And laying AI AI on top and expecting a miracle is just delusional. But set that aside. The one that sort of satisfies me somewhat, and this is perverse, I know, but you'll have a company that will lay off, let's say, 40% of its development staff because they know they can have the AI do that. That junior software engineering role. Okay, fine. I won't argue or quibble that actually they're right. What they're finding out, though, and this is the funny part, they're finding out that, yeah, they paid, let's say they paid that junior developer $100,000 a year, Mindy. Well, they now have to pay Claude or OpenAI that $100,000 a year, $150. $200,000 a year in token costs. So Mindy

Corbett: Yeah, mm-hmm.

Cale: yeah, right? Okay. So Mindy

Corbett: Right.

Cale: what happens when the token bill comes due and they didn't budget for it? Or the Mindy

Corbett: Right.

Cale: person who's who's who's in charge of PL says, whoa, whoa, whoa, wait a minute. What's the where'd my where'd all my money go? Okay. Who's who's Claude? Why is he taking all my money, right? So they'll put a stop, and this this literally happens, they'll put a stop on that payment, right? Okay. Mindy

Corbett: Mm-hmm.

Cale: Now of course, you know, when you freak my system because you cut off my claude, I'm gonna go screaming, you you gotta give my claude back. All right. So now you've got this conversation of round where'd all that money go? What's my ROI? What did I get from my Claude money? Number one. Number two, here's the really funny part. You don't pay Claude, he just stops working. Mindy

Corbett: Yeah.

Cale: Right? Okay. You don't pay open AI. He just stops working. Just Mindy

Corbett: Yeah.

Cale: boom, you're done. You are now, quote, held hostage by Mindy

Corbett: Right.

Cale: that machine, that company. They do not have Mindy

Corbett: Right.

Cale: to do anything. You've given them your data, you've given them your logic. If you even understand it, you are now held hostage till you pay the bill. At the very least, if you had that hundred thousand dollar employee that you still had on the payroll, you could ask them for overtime. You could ask them Mindy

Corbett: Right.

Cale: to please. Tell me where the logic is, what's going on, what's why is it broken? And they are theoretically not going to to go on strike and hold you hostage like the machine will. So I Mindy

Corbett: Mm-hmm.

Cale: giggle a little bit as a human being going, ha ha ha, you know, unintended consequences, you know, have come home to roost, right? now, Mindy

Corbett: Yeah. Right.

Cale: so companies, some are smarter than others. They predicted this and saw this coming and they planned for it. They didn't just go chasing AI for a a marketing gag or to get out ahead of the competition and say that they were native, you know, AI native or AI first or whatever the case may be. those companies have my respect. And they're the ones that I want to have as customers or work with, work for. It's Mindy

Corbett: Right.

Cale: those ones who were chasing that that label. And now we're getting some of their own some of their own comeuppance. I'm like, good. all that Gen X snark aside, here's the here's the the happy thing for me. You know, if if I were trying to look five years ahead like I asked you to, these lessons we're learning and these pain, these this this self-inflicted pain that we are putting on ourselves, we're going to come out of the end better for it. We're going to see the benefit of governance. We're going to see the benefit of strategy. We're going to see the benefit of not just putting a robot in charge of crappy strategy. You know, there will Mindy

Corbett: Mm-hmm.

Cale: be, if we're smart, there will be some benefits to this cycle we find ourselves in, right? But it's going to take humans to wrap this thing up, like you were saying, it's going to take these humans being involved from the beginning, middle, and end from here on out. If we're going to Mindy

Corbett: Yeah.

Cale: have that, that that outcome you and I have been each working toward. you know. Mindy

Corbett: No.

Cale: It happens every day as a new permutation, a new twist, right? A new tool. Just when just when you and I start to get a sense of how to do sort of agentry and multimodal and all that. Now we're doing graphs and loops and all those other. It's like Mindy

Corbett: Mm-hmm. Mm-hmm.

Cale: but it's fun. I will say that. Mindy

Corbett: Yeah. I will tell y'all a story. So Kale actually had drag me

Cale: Yeah, ta tack on a story. Mindy

Corbett: kicking and screaming into any type of automation like this, and that's a fact. So it all started with putting making me talk to our lights.

Cale: yeah, yeah, yeah. Mindy

Corbett: And I was like, I am not talking to my light. I'm not doing it. Why can't I just flip a switch? This is so stupid. I am not telling Alexa or Google or anything else to turn on my lights. Now, if they don't work, I'm like, why can't I turn on my lights? Because I

Cale: Like, let me try it right now. Hey Google, turn Mindy's office. not hear me. Yeah. Yeah. So I was gonna play. Mindy

Corbett: Because now I can't figure out how to flip a switch anymore. After he

Cale: Right, right. Mindy

Corbett: after yeah, after dragging me quicking, kicking and screaming. Now, what actually converted me was finding out when we had Alexa, was that I could ask Alexa, I could say, I could and we don't have her anymore. We have Google now, but I could actually say.

Cale: Okay. Mindy

Corbett: Alexa Earl Grey hot

Cale: Huh. Mindy

Corbett: She would say the replicators on this starship are not programmed yet.

Cale: This is true. When the when the Trekker when the Trekker geek and my wife found out that the the the the script, five, six, seven years ago knew Mindy

Corbett: Yeah.

Cale: it's Captain Picard. She's like, Okay, I'm in. Let's go. I'm AI native, baby. Let's do it. You know. That's all it took. That's it. That's all it took. Mindy

Corbett: Yep. Yep. I'm in I'm in now. I'm in. I'll talk to my lights.

Cale: your OG art, the other arm. there you go. Yeah, you got a Mindy

Corbett: And get get

Cale: That's too late. Yeah. Yeah, Mindy

Corbett: my Ray.

Cale: there she's she has a tattoo. There she there you go. But she's got a she's Mindy

Corbett: Mm-hmm. Hang on.

Cale: has her nice just like I do. She's got a Star Trek symbol and hers says make it so, you know, and mine says, you know, fascinating. So, you know, we are we are dyed in the wool or dyed in the ink, you know, trekkers here. So when she says Earl Grey Hot, she you know, she's OG. So that's who you're dealing with. All right. Well Mindy

Corbett: Well,

Cale: I think we have survived another episode. Whether the 17 people who listen to it do will Mindy

Corbett: Yeah.

Cale: remains to be seen. but we do thank you for sticking through it with us. Episode five coming up Mindy

Corbett: Well we

Cale: in weekish, you know, I like the word ish. Yeah. Mindy

Corbett: Yeah for this. Mm-hmm.

Cale: surprise that might make it kind of Mindy

Corbett: Yeah, yeah. I think

Cale: But in the meantime, any last words on any last words on your part? Mindy

Corbett: No no.

Cale: I'm gonna throw out one thing. Mindy

Corbett: Yeah.

Cale: So I'm gonna ask on behalf of the fan. his name is Jeff. I'm gonna ask on behalf of Jeff. sometime in the next couple of episodes, we really do need to see you play the guitar that's sitting behind you. We want to show that off, and we want to have you maybe, you know, strum a little smoke on the water or you know, whatever. whatever you want, you know. Mindy

Corbett: I play?

Cale: But yeah, you can talk about the guitar that's singing that case behind you and and maybe wow us with some of your, you know, musical talents or something. But anyway, that's it. Yeah. Mindy

Corbett: Well yeah, I consider that.

Cale: All right. Okay, thank you Mindy

Corbett: Okay.

Cale: very much, and we'll see y'all next time. Mindy

Corbett: All right.

Cale: Alrighty.