This week on The One Thing podcast, our hosts, Robin Bettenhausen and Tom Kallai, sat down with John Pingo, CEO of the Goldie Floberg, for a fascinating conversation that blended two worlds: the IDD (Intellectual and Developmental Disabilities) space and tabletop gaming.
John, a self-proclaimed Dungeons & Dragons enthusiast, provided a unique perspective on the staffing challenges in IDD care by imagining how a D&D wizard could magically solve these issues.
“Wizards can use their spells to summon more beings and creatures,” John says. “So then that would solve all of our staffing problems right there. Bring a minion – solve the staffing crisis!”
The discussion quickly turned serious as John highlighted the significant challenges facing the IDD field, particularly around information overload and the burdens placed on Direct Support Professionals (DSPs). He revealed that DSPs are expected to manage over 1,300 pages of documentation across multiple homes and individuals – a staggering amount for anyone to process. This leads to both burnout and turnover, two persistent problems in the field.
The real highlight of the episode, however, was the unveiling of a new AI-driven tool that John’s team at the Goldie Floberg has been developing: the My Life app. This innovative tool leverages large language models, like ChatGPT, to streamline access to critical information DSPs need in real-time. The app is designed to reduce the cognitive load on staff by allowing them to quickly retrieve information about individuals they serve, such as behavior support plans or medication reminders, through simple voice or text commands.
“What we’re hoping is that better knowledge, easier access to knowledge, will lead to better care, better supports, better services,” John says. “We also think it might help with turnover, which is a real challenge across the nation. Shortening the learning curve and increasing comfort level in helping the people we serve – that’s the goal.”
Tune into this episode for a compelling look at how technology can solve complex problems in the care industry. You’ll also learn more about how John and his team are bringing the future of care to the present, one app at a time.
TRANSCRIPTION
Robin Bettenhausen (00:03)
Hello and welcome to the Horton Group’s “The One Thing” podcast. We are so excited to be here with you again today. I’m Robin Bettenhausen. I’m here with my colleague, Tom Kallai and our special guest of the day, John Pingo of Goldie Floberg.
John Pingo (00:19)
Hi, great to be here, thank you.
Robin Bettenhausen (00:21)
Thanks for joining us.
Tom Kallai (00:22)
Hey John, so I’m always excited to see you because I get to pick your brain about something that I’m very interested in and I know that you’re interested in it and every time we talk about it, you talk about it with excitement and passion. And while the crux of our conversation today is gonna be about the IDD space, this is a different space. It’s tabletop gaming. So I’m gonna bring it back around to the IDD space though. So if…
If you had one position, okay, at Goldie Floberg, that’s open right now that you could fill, but you had to fill it with a character class, okay, so for those who don’t know, the uninitiated character class is like a fighter, a thief, a bard, all, they have very specific attributes to them. What position would you fill and with what character class?
John Pingo (01:14)
man, that is a fantastic question for a super nerd like me to start out with and I know the answer right away. It would be a wizard from Dungeons and Dragons and it would be a wizard specifically because they can use their spells to summon more beings and creatures. So then that solves all of our staffing problems right there. bring a minion, that’s like the… Solve the staffing crisis. You bring in a wizard and that is like the ultimate…
Tom Kallai (01:19)
you
Okay.
Robin Bettenhausen (01:32)
I mean…
Tom Kallai (01:34)
That’s right, we’ve solved the snapping crisis.
John Pingo (01:41)
human resources manager right there. They’re like casting spells and summoning all the health you need. So that would be perfect.
Tom Kallai (01:48)
Perfect answer, perfect segue to question our first question here, which is, you’re in a good spot, a high spot at Goldie Floberg where you can kind of see the lay of the land. So what are some of the biggest challenges that you see facing the IDD space?
Robin Bettenhausen (01:49)
Love it.
John Pingo (02:05)
Well, I think that’s a really good question. are multiple challenges, the perpetual challenge of funding and making sure that direct support professionals who are just the absolute backbone of all of our organizations have a professional living wage and pathways for advancement. But another big issue, I think, facing our field and specifically our direct support professionals is
one of information and information overload. We live in a, and work in a system that is just continually growing more complex and our direct support professionals are expected to know and need to know so much information about the people we serve, about the regulations that guide our organizations, about our own internal organizational policies and procedures. And just, you know, just to kind of…
help put some numbers on this. was looking at for a direct support professional working just for one person served. They need to know on average 1 ,312 pages of information and that’s just for one person. I’m sorry that’s for they need 1 ,300. Sorry I screwed that up. So
When you look at the amount of information a direct support professional needs to know, I just went and did some back of the envelope math a while back. And you look at just the information related to one person served. In our organization, that’s around 110 pages of information. And we’re not talking big, giant 24 point plot. We’re talking like 10 point plot.
Then you throw in our employee handbook, which has all of our critical policies. That’s another 206 pages. And then you throw in our operations manual. That’s like all the job responsibilities. Like how do you actually execute your responsibilities in a safe and correct manner? That’s another 226. So we’re talking about 1 ,312 pages of information because you have to think each direct support professional isn’t responsible for one person’s
Like in our organization, they’re responsible for a minimum of eight because they get cross -trained at different homes. We operate 20 different community group homes and soon a high intensity support program for individuals with developmental disabilities. So again, that’s over 1 ,300 pages of information that a direct support professional is expected to not only know, but know to a level that they could pull it out of their mind and act on it.
instantaneously. And I don’t think that’s a realistic expectation for any human being at all. And I think that’s a really big challenge because that information, that level of complexity, is going to increase over time. It’s not going to decrease. Our world’s growing more complex, not less.
Revolutionizing Support: AI Tools for Group Homes
Robin Bettenhausen (05:13)
Yeah, gosh, it feels like maybe there should be a tool to help with that, right? And I think that’s what brings us to our conversation today. Tom and I each have three kids. And so as I’m thinking about what you’re saying, like how to know that much information about any one human, even one we live with 24 seven is challenging, right? I mean, I always make the joke of my kids one week love yogurt. And so I buy one and then I buy one more.
Then I buy 12, right? It’s a better discount. And suddenly everyone hates yogurt. So it’s constantly changing information and how do we best manage that? So we have heard that yourself as well as your organization have been working on an app -based artificial intelligence tool or, you know, sort of resource that you’re going to be deploying into your group homes and other programs that you have. So tell us a little bit about how that came to be.
Tom Kallai (05:45)
you
Robin Bettenhausen (06:09)
the problem it is that you’re solving for.
John Pingo (06:12)
Yeah, we are super excited about this. And it all goes back to that problem, that conundrum of the massive amount of information direct support professionals need to know and be able to act on in regards to the people we serve. mean, first, I don’t want to come off as like one of those like AI CEOs who are just like hype machines, because there is so much hype with AI right now. It’s going to create a pulse -scarcity economy. It’s going to cure all disease and solve climate crisis.
You know, no it’s not. But if you push past the hype, and specifically when we’re talking about artificial intelligence, what we’re looking at in our organization are large language models like chat, GPT, Gemini, Lama. These are actually really powerful tools for problems that involve organizing and accessing really large amounts of information in a very user -friendly manner. again, this,
Tom Kallai (06:44)
you
Robin Bettenhausen (07:07)
you
John Pingo (07:11)
problem of DSPs needing to know and be able to access this large corpus of information about all the people we serve is actually ideally suited for these large language models. Now, for example, we’ve got a cloud -based data management system that contains a tremendous amount of rich information about all of the people we serve. But I was looking at just
just to access a very important document, what we call a behavior support plan. And that has all the behavior support strategies for the individual on how to help them meet their needs so they don’t ever need to exhibit contextually challenging behavior, but also what to do when those sort of behavioral situations arise. A really, really important issue. In our current data management system, which is an excellent system, it takes…
five clicks and typing a name to get to a PDF of the behavior support plan. And then you need to read through the PDF. Now, I don’t know about you, but I was a direct support professional and I know from experience, I would not in a crisis situation, I would not have time to do all of that. The current system has all the information there, but it’s too much friction. It’s too much of a response effort to access all of that.
important information. what we are doing is we’re developing an application that we call My Life. And it’s Mi -Life. And it’s My Life with an I. I was spelled with I because I is the most person -centered letter in the alphabet. Also, my with a Y. It was already taken. We couldn’t get the web domain. So that’s the really like
philosophical reason for going with the I and then another reason that’s not exactly right. So what we’re doing with my life is we’re taking a large language model in this case is cat TTT 4 -0 and we’re inputting into it all of the information that we have about the people
Robin Bettenhausen (09:12)
The old practical reason.
Tom Kallai (09:13)
We won’t say which one came first, right?
Revolutionizing Support for Professionals with a User-Friendly App
John Pingo (09:37)
We serve and what this is going to and what it is actually allowing us to be able to do then is there’s an application that will sit on our smartphones and our tablets that we have in our homes and staff will be able to tap on the My Life application. You sign in once during your shift and then you’re signed in for the whole rest of the shift so then you just tap on the application.
It’ll pull up the individuals that you specifically serve. It’s just a nice little portrait picture and their name. You tap on that picture and then you can start asking it whatever questions you have and you’ll receive an answer both in text and spoken language format. So it’ll tell you the answer and then it’ll also put it in text so can scroll back and reread it as you need. And so at its most basic function, that’s what the application is.
And what our hope is and what our belief is, there’s such a huge learning curve right now for direct support professionals to learn everything they need to know for the people they serve, that this is going to dramatically shorten that learning curve. And what we wanted was an application that was as easy to use as texting on your phone. And because if it’s that easy, people are going to use it.
And that’s what we have. It’s no more challenging than texting. It’s actually a little bit easier because when the application provides you the information, it will automatically speak and tell you the answer. So don’t even have to, if you are pretty busy with something, you can just listen to it. You don’t even have to read it. But if you want to go back and review what my life just told you, you can go back and do that. we’re really excited.
simple idea that solves a pretty complex problem, or we think it’s going to solve a pretty complex and challenging problem. We’ve been starting to work with our staff to kind of get them educated about it. We’re actually rolling out the application in our first home in October here, very shortly. So we’ve been working with staff to kind of prep them for it as they field test it. So we’re really excited in terms of our timeline.
goal is to have it field tested in first home and then have it rolled out to all of our homes in the first quarter of 2025. And so we’re moving really quick. I just met with the development team working with a fantastic developer called Top Flight. And we met with them. We have a weekly meeting. We met with them this morning. And right now I have a prototype version. We have a prototype version on our phones, but it’s all text -based.
right now. we, the, the, the chat bot, the knowledge base for the first three people served is all built, but it’s all text based and they were, testing and showing us the first, verbal like language component, this week. So that was really exciting to have the app start talking back. And the end goal will be that the app takes advantage of these large language models. what’s quickly evolving is just conversational speed.
It’s getting really close that, and we’re talking like really close, we’re not talking years, we’re talking months to a year, and these models will be fully capable of having fluid conversational speech. And I think that’s really the end goal in terms of the user interface with MyLine.
Tom Kallai (13:22)
think there’s one thing that you said that resonated with me there, which was that there’s just too much friction currently, right? And that can be kind of like a layered thought, meaning like, A, there’s too much friction just on the information you need to do the job baseline, right? There’s a ton of information on top of that, if someone does have like a behavior response plan. There’s some nuance to understanding what that is. And there’s also a ton of data that has been taken usually to back up what we’re going to execute on if a behavior.
John Pingo (13:28)
Yep.
Yes.
Tom Kallai (13:51)
is presenting. So do you see this is kind of like bringing down that learning curve maybe on the on that I always call it kind of like a sixth sense that when you work with people, you kind of get to know, okay, they’re at they’re elevated, maybe reduce some friction in terms of understanding the person and and like the job.
John Pingo (14:15)
Most definitely, Tom, I think you really nailed a key hope we have for this application. It’s not just getting the technical information to the direct support professional in the easiest manner possible, but it’s also about improving quality of service. we believe that better knowledge, easier access to knowledge will lead to better care, better supports, better services.
You know, it’s going to be really interesting. We track data on turnover and everything like that. So I’m really interested to see if we see any sort of improvements or changes in terms of direct support professional turnover. That’s a real challenge in our organization and in many providers across the nation. And that actually makes the whole learning curve issue even more challenging because you’re starting over frequently with direct support professionals. So anything that we can do to shorten that learning curve
to help a person, direct support professional become an expert on the people we serve. And increase, as you’re getting at time, increasing their comfort level in helping the person. And what we’re hoping is that if you have this application that is so easy to use and you can ask it whatever questions you have about people served, that’s going to increase your comfort level. And you’re more likely to do it because it’s just a tap and then you’re talking to it.
Tom Kallai (15:36)
Yeah, it’s right. It’s easy. It’s right there. It’s in my hand. It can talk to me if my hands are full. I can talk to it and it can talk back to me rather than have to click five times to find and then read and then find it. It’s almost like you found a wizard and some spells to take care of some problems.
John Pingo (15:58)
You know, all those years of playing like Dungeons and Dragons since I was like 11 years old, they finally paid off.
Tom Kallai (16:01)
There you go, yes.
Robin Bettenhausen (16:03)
You mentioned you’ve been piloting the program. What’s been the feedback so far from your staff members?
John Pingo (16:12)
So right now we have like an internal development team and we’ve been piloting it and playing with it. so we have three people served and we’re working with their families. So we’ve got their initial chat bots developed and we have, created a chat bot for myself that I can then demo and show to people without, you know, without worries about, you know, confidentiality or anything like that.
I think staff have been pretty impressed with the quality of information that they get and the depth of questions they can ask because again, these large language models are so good at looking through huge amounts of data and taking your question and then finding what you need. And then you’re able to refine your questions. For example, if we go back to the behavior support plan example,
These language models are able to really clearly understand and parse out information if you’re asking just a general question like, hey, can you tell me what John’s target behaviors that we’re trying to help him with are versus can you tell me what to do if John starts hurting himself? What should I do? And it’ll find that specific information in the behavior support plan and just pull that information.
for you. Now, granted, the information is only as good as the information we give the model. So that’s one of the things too, is it’s really pushing us as an organization to go back and make sure that we really have just the best quality, most in -depth information that we can put together on each person served. But yeah, the teams have, having a lot of fun playing with just the…
the text base app and it’s easy to use and it’s just going to get easier once the verbal component, the language component comes in next month. And we’ve also put together like a whole series of like test questions, you know, to kind of ask it like, here’s like the bare minimum, like questions that we think direct support questions are frequently gonna wanna know. And we, you know, as we bring each person’s, you know, knowledge base and chat bot online.
we’re going through those questions to make sure it answers all of those accurately. And that’s another cool function that’s being built into the MyLife application is people will be able to talk to MyLife and say, hey, can you add this information to the person’s knowledge base? I just discovered that John really likes this show on YouTube.
Can you add that in there so that people know to help him, you know, watch the show because it’s one of his favorite. And then what it does is that gets flagged and that request goes to the person’s case manager here and going forward. And then they get to act as the gatekeeper to make sure, yeah, that’s information that should be added to the knowledge base and tells the AI, yes, it’s okay to add this to the knowledge base on individuals. So there’s kind like little safety mechanism to make sure we get.
good quality information. getting back to, know, like a service plan should be constantly evolving. And this gets us to a point where it could be updated and the knowledge base on the person could be updated multiple times a day if that need arises. So we’re really excited about that.
Robin Bettenhausen (19:57)
Bye.
Tom Kallai (20:03)
There’s so many cool ways that this can go. I think of like, you know, people talk about AI and they think like the end of that is like the singularity, which you’ve heard of where everybody’s uploaded into it. But this is kind of that idea where it’s like, think about all the knowledge that you lose when a DSP leaves, right? They had like intimate knowledge of that house, how it ran, favorite shows, all that stuff, like at the ready. If you can make that at the ready to some other person who’s coming in, it’d be very interesting to see.
John Pingo (20:12)
you
Tom Kallai (20:33)
the effects and it’ll be very interesting to see what gets tracked too, right? Out of it, like have our behaviors come down, you know, injuries, have injuries gone down, have medication errors come down because everybody can just pull up exactly what it’s supposed to be at any time. Very, very cool.
Enhancing Care with Mi-Life App Features
John Pingo (20:41)
Yes.
And Tom, that reminds me of another function that’s getting built in to the MyLife app that’ll be up before the end of this fiscal year is you hit on something that is as our internal development team was meeting with our app developers that we felt was very important is the ability to also for the application, MyLife application, send push notifications of important reminders to our staff. And one of them is, Hey, don’t forget it’s 1130 and John takes such and such medication at noon.
Tom Kallai (21:22)
Right.
Robin Bettenhausen (21:22)
Mmm.
John Pingo (21:22)
And that gets back to just the trying to reduce the cognitive burden on our direct support professionals because there’s so much stuff they have to remember. And a lot of the stuff is changing frequently. It’s like, John has strep throat, so he has an antibiotic. That’s not a regular medication you get, so it’s easy to forget. So we put it, we tell my life, hey, please remind all the person who works with this individual served.
Tom Kallai (21:28)
Yeah.
John Pingo (21:51)
that he has an antibiotic and he gets it at four o ‘clock. So please remind them of that. like, sign, a direct support professional signs on in the morning and they get a notification right away, hey, don’t forget, John has a dentist appointment at nine o ‘clock this morning and here’s some things you need to take with them. And, you know, in our organization, there’s a pretty involved process for staff to become like approved drivers through, you know, the state and everything like that. So we have a limited number of drivers.
Tom Kallai (22:07)
Yeah.
John Pingo (22:21)
And so sometimes the direct support question on taking someone say to a doctor’s appointment may not be the person who’s most knowledgeable about those individuals. So now they’re going to be able to go to the doctor and the doctor or the nurse is asking questions. They’re going to be able to turn on My Life and say, let’s just ask My Life. And if it’s medical information that we know about that person, that application is going to, the My Life application is going to be able to answer all of their questions. reducing, improving the quality of that experience.
Tom Kallai (22:37)
you
At the ready.
John Pingo (22:51)
for the person served and reducing like the stress levels for the direct support professionals. Like, I don’t know, let me try and call the nurses and see if I can get one online. And then the nurses are digging through all the medical stuff and everything like that. there’s other kind of like a ripple effect that we think this application could have in terms of this information will be helpful for direct support professionals. We’re looking at if parents are interested, they could put it on their phone or their tablet and you know, they would obviously have access to their loved ones.
Robin Bettenhausen (23:06)
Great.
John Pingo (23:21)
account, but think about getting like real time, just updated information about who knows a loved one more than mom or dad or brother or sister and stuff like that. And having them be able to not only check the information and let us know, Hey, this is accurate or this isn’t, but also updating it. And, you know, if, if they need to take their loved one to a doctor or something like that, it’s like, I’m not up on like this, you know, this week’s medication change or
Tom Kallai (23:29)
Right.
Robin Bettenhausen (23:29)
Yeah.
you
John Pingo (23:49)
or when’s the last time they had their dentist appointment or something like that, they can pull up the application and say, let me ask them, like that and get that information right away.
Robin Bettenhausen (23:59)
It’s such a, I know you said it was like a simple thing, but it solves such a huge problem. And isn’t that really what we look for? We look for a pretty simple solution that, like you said, just really removes the friction. I’m thinking, you you probably had a whiteboard in the house. You probably had notes being left on the counter as shifts are changing and just taking that out and using technology in a way that hopefully it was intended to help.
the individual have the best possible experience and to help your staff members have a better experience. I mean, in my head, I’m just thinking about a page, a stack of a thousand pages of stuff, right? Nobody wants that. Nobody. So if I’m thinking about who’s listening to this podcast and I’m guessing it’s groups that are other potentially IDD agencies or others that maybe are serving individuals and they’re probably thinking as I would be, how can I get my hands on this or what’s next?
So tell us about what your end goal is or sort of next steps in terms of distribution. know pilot program rollout. Tell us about that.
John Pingo (25:02)
Yeah, again, we’re really excited about it. So this first fiscal year, what we’re working on is just getting all the core application functions up and running and just working really well within our organization. So that takes us to the end of this current fiscal year. Part of that is also we’re working with, for example, Thera, which is a major cloud -based data and document management
service and an excellent service in our field. We’re working with them to use their API to interface directly with their app. So my life would be able to pull information out of their app. And as it’s updated, the information would be updated, which would be huge if we’re able to work with their app and other sort of like large information management systems in our fields.
That’ll be one of the next big kind of functionality pieces that we add in but What we’re going to do is we’re actually going to spin my life off as its own separate Corporation and the wheels have already started turning on that process with the goal being in the next Probably 12 months having it spun off as its own company and
with the possibility of bringing on other interested providers as investors and owners of the My Life application and what will hopefully be eventually a whole suite of applications that could improve the lives of the people we serve and the people who serve them. So we’re really, really excited to kind of take that also to the next level. you know, we’re hoping that some of our
colleagues out there might be interested in joining with us because we think this has a lot of utility for us. It might have a lot of utility for other providers who might be struggling with kind of the same sort of information overload issues. But I do think these large language models also have potential to solve some other problems or at least lessen some other problems that we struggle with.
internally just starting to have like discussions trying to identify, okay, here’s what these large language models do really well. Where are some other areas that that might be able to help? And maybe it’s not, you know, maybe, you know, this one is DSPs, maybe the next application might focus on case managers. We’re just, those are like real initial discussions, but that would be another benefit of spinning this off as its own company to get a, you know, a group of providers with all of their varied experiences.
talking about, how else can we apply this technology that already exists and to improve our services.
Robin Bettenhausen (28:02)
you
Yeah, it just feels like there’s so many opportunities there, right? I mean, where you’re developing it is out of a need, but you think about anywhere there’s large groups of people with a need to know a lot of information, whether it be a school, whether it be a correctional facility. I mean, the list goes on and on and on where you’re just around a lot of people that requires a lot of data to be known. just think.
this is going to be an awesome application and solution. And we’re excited to see where it goes. We’re so thankful that you joined us for this podcast today. And we’ll be sure to include some information for people that may want to contact you. know Tom and I will be signing on as investors if you give us the chance. So keep us posted on that.
John Pingo (28:49)
I really appreciate the opportunity Robin and Tom. It was super fun talking with you. Thanks Tom for the D &D questions. Those are some of my absolute favorite favorite questions. The answer I almost never get them. So I really appreciate it. This was a blast. Thank you for having me. Really appreciate it.
Robin Bettenhausen (28:57)
Thank
Tom Kallai (28:57)
Anytime.
Robin Bettenhausen (29:02)
You
Thank you so much and maybe we’ll do an update next year on how it’s going and what’s changed and what you learned. That would be really fun.
John Pingo (29:16)
That sounds great. Let’s do it.
Robin Bettenhausen (29:18)
Awesome, thanks so much, John. And thanks to our audience for joining us for The Horton Group, the One Thing podcast. See you next time.
Material posted on this website is for informational purposes only and does not constitute a legal opinion or medical advice. Contact your legal representative or medical professional for information specific to your legal or medical needs.
