Miriam Allred (00:00)
Welcome to the Home Care Strategy Lab. I'm your host, Miriam Allred. Today in the lab, I'm joined by Matt Erickson, the VP of Sales and Operations at Griswold Corporate. Matt, welcome to the show.

Matt Ericksen (00:12)
Hey Miriam, thanks for having me. It's great to be here.

Miriam Allred (00:15)
I have been looking forward to this for a while. You and I have been friends for a long time. You and I have been in home care for a long time and I've interviewed you several times and it's time to have you back in the lab on the new show to dive into AI. I am just so impressed with your passion and excitement for home care, but also your passion and fire and what sound like sleepless nights thinking about AI and its implications and its use cases in home care. And so I...

want this to be pretty like casual, conversational. I want to pick your brain on what you're doing, what you're thinking about, what are you implementing with your franchisees? I want to just kind of take you in a few different directions and think through and talk about this in a more just casual conversational way. So before we get into it, ⁓ for those that don't know you, why don't you share a little bit about your background in home care?

and then bring us up to speed on your current role and responsibilities at Griswold.

Matt Ericksen (01:14)
Sure. I've been in home care and home health care now for coming on 13 years and I've held every role from scheduling coordinator, recruiter, outside sales to location director, everything but clinical pretty much. And so it's been an amazing journey that allowed me to see all the different aspects of the business and what goes into those roles and to be successful and a lot of potholes along the way too.

But that's led to the wonderful place that I'm at right now with Griswold Corporate, the franchiser based here in Bluebell. I've been here coming up on five years now. ⁓ in my role, taking all the life experiences I've had on the operator side, working on the franchisees and their corporate structures, pediatric home health, adult geriatrics, behavioral health.

⁓ is bundling all those together to best serve our franchise owners, their teams across the country, and most importantly, their caregivers and the clients that they serve in their communities.

Miriam Allred (02:18)
I love it. And I might cite this at the start and several times in the conversation. You got a degree in history. And I just think it's funny and interesting how we all came to home care. But history degree turned home care, turned VP, turned we're going to talk about it. And you're going to show some of your AI tools. now you're coding AI like the rest of the world. But it's just amazing how we all evolve and where we land and what we're capable of now in this modern AI world.

Matt Ericksen (02:45)
It's

yeah, it's absolutely it's like I couldn't write a crazier script and what they kind of like where my life's gone with with with this I have an amazing amount of random facts in the Civil War if everybody ever wants to know those that's that's the specialty of my degree but ⁓ ironically what I found in my as I got immersed into generative AI and LLMs is it's not the typical IT ⁓

specialties that these companies are hiring for. They're actually hiring individuals with degrees in English, history, political science, ⁓ more of that aspect for the ability to spot trends and analyze and prompt engineer. And I thought that was just kind of cool. was like, who knew that a history degree could actually be useful in the skills that are now applicable to the world that we live in.

Miriam Allred (03:34)
Yes, and before this conversation, I think you were listening to a podcast this morning talking about like scaling up like in our career in our life. It's all about like scaling up. But with AI, it's totally like rewriting that script on how we think about like accelerating our career. Share a little bit more about like your thoughts on what you heard.

Matt Ericksen (03:53)
Oh, it was, it was

one of those light bulb moments that you have that you wish you're always chasing you when you have one. just like, Oh, that's so profound. And I was just, I was listening to a, there's a practical AI podcast and name escapes me of the exact title of it, but they were, they were emphasizing that

you know, historically we've in the workspace and workforce development, we've always upskilled in vertical channels, you know, taking a coordinator to maybe being a senior coordinator or director of an office, right? You know, and just up and just kind of keep it on a specific track. Well, you know, it's not about upskilling anymore. You don't have to make a better say coder or you have to make a better ⁓ scheduler in a way as much as you have to help them access and learn how to just take advantage of the tools.

take advantage of prompt engineering. There's a big difference between just searching Google and structuring the correct prompt and teaching somebody how to do that. You can create quite the multiple on their skill set very, very quickly. And I think just as a material shift from just historically taking years to learn a trade, that's not the case anymore.

Miriam Allred (05:01)
It's so interesting the multiply. I completely agree. It's how do we take a role or a function and multiply it with AI? I think that's the realization that a lot of people, companies are having is how do we, how do we like double down and multiply with AI? And that's where things get really interesting. And that'll be a tone through the conversation today. I want to take you back, I don't know, a couple of years, like you've been tinkering with AI for years now. I want to hear your first aha moment.

that you were tinkering, you built something or did something or thought something and it like unlocked this like this passion or this fire in you of like, wow, this is really gonna like change the game for home care. What was that moment for you a couple of years ago?

Matt Ericksen (05:47)
Yeah,

was definitely, it was somewhere and it's all blurry now because days just fly by. But it was somewhere between like, I would say mid to late summer of 2022 and, your GPT is coming out and we're all using kind of like as a glorified Google search or again, at that point it was a lot of, can you help me write an email? Right. And

That was about kind of where everybody's kind of public use, general knowledge base was, and certainly, it helped me there a lot. I struggle with putting eloquent emails together at times, so that it certainly brought value, and that kind of, I'm like, okay, it can do this for me, what else can it do? ⁓ I've been heavily involved with the caregiver and staff recruitment retention strategies here during my tenure, and you know.

Going back to our indeed page, know, every everybody's got a company page. Everybody's got reviews. I'm like, well, what can I do with that? That's a giant data set that I know I can upload this file to GPT. What can we do here? And that's a Part of the fun of it is just kind of like innocent exploration, you know I don't have any intention of kind of figuring things out. Hey, what does this button do quite literally and I downloaded all of our ⁓ all of our caregiver reviews I

upload the file to chat GPT had, you know, at that point it had the innovation to do analytical capabilities, right? And it could look at numbers and analyze them and quantify stuff. And it was able to work. I was able to prompt it to not knowing what I was doing. I have no clue what I was doing. So I'm not going to act like I was just by accident, right? But that was the aha moment was I figured out like, oh my God, I can organize by region, by state, by location.

sentiment analysis, a behavioral analysis, a strength and weakness analysis of our franchise network and I can take this information that I created in 15 minutes and now I can go and equip our owners with insights that would have taken days to historically put together in a manual process. Even if it was done in Excel, it would still have taken so long. ⁓ that was the aha. like the the the aha was not so much that this is cool that I can just go and do this like

quick analysis, it's the thing that had me sit back in my seat was I just did in three days, what did it take three days? I just did it in 15 minutes. What else can I do in that compression time? And then that from there, that's where my wife would come downstairs at 2 a.m. in the morning, why are you still up? I'm like, because I'm fiddling and I'm tinkering and I'm like, but that was what set off what I do now.

Miriam Allred (08:29)
And that was three years ago. And I think what you just shared might be mind blowing to some people. Like, wow, like that's already like something that other people can go and replicate. But it just gets me excited about what you're going to show here in a few minutes. It's like insane how far you've come, how far this technology has come in three years. And that was an all how for you a couple of years ago. And you just like, you know, hit the fast forward button and dove into this like head first and are seeing like insane results.

results with this today. one more question just to kind of like tee up this conversation. actually two. One is ⁓ it took home care a while to adapt from paper to software for obvious reasons. What's your take on our transition from software to everyday AI adoption and use? Like what's your take on what that time span is going to look like?

Matt Ericksen (09:27)
I think it's gonna compound is the best way I describe it to people is that for the innovations that are occurring in present state and that we know are only a matter of time, it's only gonna compound and just speed up. And when we look at, take a step back from the environment that we're in right now with everything and the big jump from paper based to then going to our digital.

you know, CRM based kind of style that we're in right now. You know, that's a material shift in, you know, in the work environment and the, you know, at that time, the upskilling, you know, how do we, how do we change a skill set or improve a skill set from paper based communication styles, right? Where a lot of time between tasks, right? To this year, I'm absolutely sped things up, but now we don't have to make that jump from a different environment. We're in that same digital environment. We're just changing the tools.

that we're adopting into our daily life. that's the key right there is the, ⁓ that's where the compound is gonna occur is that same environment, if you have the right talent and they learn these tools, then we're gonna be moving, what would have taken five years, we'll do in a year, two years. ⁓ I don't see a world without AI being commonplace. ⁓

if that's the best way to say it, know, for another year or so. Like it's just gonna be so visible, top line, embedded everywhere. We'll probably forget that we're actually dealing with it at that point. It's just gonna be so kind of like an unconscious interaction every day.

Miriam Allred (11:10)
Yeah, I think it'd make a really good point is we already have the digital foundation. So we're not necessarily, you know, establishing an entire new base, a new environment. I like that you're using that word, but it's just actually like layering in new opportunities, new tools, new ways of thinking, I think more than anything back to like prompting. We're not necessarily teaching people an entirely new skillset. It's actually just like leaning into, like language.

and writing and prompting. ⁓ I'm curious today, think of like your network of franchisees, what the AI penetration or adoption is. Like just share like your take on how many people are tinkering, how many people are using it and seeing results, how many people are not using it, how many people are like, you know, keeping at an arm's length from this. Like what are you seeing across your network?

Matt Ericksen (12:03)
Yeah,

yeah, ⁓

Whatever I see on a daily basis, whenever we interact with it, it always impresses me because I think that's beautiful thing about ⁓ the space we're in with this AI innovation chapter is not one single person has all the best ideas or they have all the bright ideas, right? And so I get excited about talking with franchise owners because each one of them is finding different ways to find success in their own way with these tools. it's just so exciting and just...

it's a great time to be at Griswold for that reason, is that the owners are just all finding ways to adopt it. to answer your question, so I don't go down too far of a rabbit hole, is ⁓ after that aha moment with the indeed experience, I was like, okay.

This is a thing that's not going away. We have a conference coming up. How do I get us positioned well with this to be commonplace in our brand? And so when I was on stage doing one of our presentations on AI and what it means, because everybody's asking, well, what is this thing? What is this thing in 2022, 2023? I asked at a show of hands, how many people have played with this thing? How many people have tried this back at the end of 2022, beginning of 2023?

and way more hands I could have ever possibly thought would ever raise went up.

And then when they started sharing examples and the mic went around and people are talking about how they were using it back then, I'm like, people get it. It assimilates almost naturally. know, just the pain points are so great with the business that we're in and what it can solve that people just kind of were like, all saw the connections from whether you were a marketer or clinician or wherever you came in your life. Like we, everybody at Griswold saw that opportunity. And at this point it's 80%.

85

% of our network is interacting with it in one way, shape, or other. ⁓ I've yet to meet somebody that's like absolutely not. Not everybody. mean, from the person that, you know, from an owner that opened up a year ago to, you know, our 10-year donors that have been with us for 35 plus years, they're all using it and they're all finding ways to make, help it, make their care delivery better.

Miriam Allred (14:23)
And the results might be slightly biased because they have a leader like yourself that's just like all in on this and you are setting I think a good example of like let's tinker, let's play, let's figure this out together. But it's interesting and I think good to hear that maybe 80 to 85 % are using it in some way, shape or form. My hope is and I do think that might be slightly reflective of the industry. think everyone I talk to, it's definitely on their radar. You everyone's thinking about it, talking about it. And I think the adoption, the usage, the tinkering has gone up significantly.

especially in this past year. And so we probably are at, you know, maybe that 80 to 85 % as an industry. And so those listening to this, if you're in that 10 or 15 % that's not in, like get in, get on, like you said earlier, there's no going back.

Matt Ericksen (15:06)
You can't break it, just why not, right?

Miriam Allred (15:10)
Why not? ⁓ Tell me the top five or six tools that you personally and professionally are using. Obviously everyone's using ChatGPT, OpenAI specifically, but there's maybe five or six tools that are in your tool belt. Rattle off what those are and maybe just briefly about what you use each of them for.

Matt Ericksen (15:28)
Yeah,

so Griswold is a Microsoft ⁓ environment. ⁓

our senior director of IT, Tim has done a great job as far as bringing all those tools and resources into one easy integrated environment. And the co-pilot from Microsoft really layers well over all that. so on a daily basis, my operations team and I, we use it in all our functionality, whether it's in Excel, whether it's in PowerPoint, whether it's in Teams. ⁓ It helps us record transcriptions and summaries. We've weaved it into every which way possible with

our daily interactions with each other with our franchise owners. ⁓ It's helped us be overall more productive there. ChatGPT, ⁓ obviously it's the most common one that everybody talks about. ⁓ I like it. It replaces Google for me, to be honest, in that it allows me to articulate a search prompt far more effectively and get faster results ⁓ than I otherwise would have to do for five.

let's say manual searches that aren't, with no intelligence behind it, right? So I like it for that. Claude from Anthropic is a great tool. Each one of these agents or each one of these LLMs have a specialty. Claude we found is really great at coding as well as helping to edit and to help you finesse emails especially. There's a high volume obviously, but it's really great as far as helping

to add a tone to it. ⁓ I found that one would be like a lot. RingCentral is our voiceover IP service, transcription services from their AI notes. If you have a voiceover IP system, ⁓ ask your provider, what's it cost? Can this be added on? Sometimes they have it on for free. It really is kind of like a...

If the voiceover IP provider doesn't have it, I'd switch providers, let's put it that way, because it's just such a game changer there. ⁓ And one of the creative ones I really liked that I actually used for that first initial presentation ⁓ on stage at conference was, it's called Dectopus, and it helps you essentially, you can prompt engineer yourself a PowerPoint. You don't have to manually create it. And so from a creative side, it's super great to... ⁓

you know, to take information, maybe you pull it from, you know, do an analysis on chat GPT, just dump that information into Dectopus and it's going to help you just kind of build a, build a presentation in a fraction of time that it does manually.

Miriam Allred (18:09)
Amazing. so prompting is like the underlying theme here. I want, I want to actually have you explain and talk a little bit about prompting or even share like the anatomy of prompting because all of these tools function very similarly. If not the same, you are basically prompting it. It gives you outputs. You prompt it again to refine those outputs. And so we're all at this stage of like getting better at prompting and thinking about prompting and having AI help us better prompt. So will you

that like anatomy of prompting and why that's so key to having success with any and all of these tools.

Matt Ericksen (18:46)
Yeah, absolutely. I'm going to share my screen.

So what I'm sharing here is this was actually, it's just a screen grab that I did when one of those nights I couldn't sleep and I stumbled upon, I think it was an Instagram post ⁓ from ⁓ OpenAI, but with the latest GPT-5 release, they also sent out what's the ideal way to structure a prompt. And I've been using this ever since and I can validate it, it works. And to walk everybody through it is, prompt engineering is, it's not complicated, it just needs to be well thought out.

It's Google, everybody say, hey, just go Google it, right? And you can ask, we've gone lazy with our requests. It's just like, we've almost been able to talk SEO. So we literally do our searches based off of what we know to be the keywords that are necessary to trigger result in SEO. We're just reconditioning ourselves now. And it requires us to not be lazy in a way where to get a really great result with these generative AI platforms.

⁓ The prompt you see in front of us, has a body to it. So it has a role, right? I like to always start my prompts off with, you are an expert in home care. You are an expert in sentiment analysis. You are a psychologist. Who do you know to be a subject matter expert in that area? And just quite simply say, impersonate them, right? Then define them. have to define the task. Be very clear. Articulate it well. Couple ⁓ dashes, bullet points.

that will highlight it, give it context. What are you trying to do here with this information? You've defined what's the role is, you've defined what its task is, what you'd like it to do, but how does that fit into the bigger scheme? Help the system understand that and then it can use its logic to make that small tweaks. And this is where the refinement really pays off. Reasoning, ask it, this is where tone comes in. How do you cross check, validate information? How does that logic go to work on providing you that?

that that output as it gets refined through this process. And how do you want to be presented it? How do you want that information once they're finished looking through it? Once the system is looking through it? Do you want it in a grid fat pattern? Do you want it in a paper based? Do you want in a word document? These are all there's no limits here. So you just got to define what how you want to be able to have this information in a viable format. And then the last and this is really important often overlooked is stop conditions is what they call is

tell it when you want it to stop thinking because it can over engineer something for you. It can over engineer a result when it may be a great idea, but it may be too many great ideas into one thing. I've stumbled, that's personally happened to me is I've created great products and tools and assets, but there's too much in one thing. I need to break it down. so tell it, stop after this point, stop after you've successfully created one insight. You want to define guardrails essentially.

This is the anatomy of a GPT-5 prompt, and this can work across other generative AI platforms. ⁓ It takes a few minutes, but it will give you results that would have taken you hours to achieve.

Miriam Allred (21:55)
This is a gold mine. I'm like embarrassed that I didn't I wasn't aware of this before you shared this with me But it's it's so simple. It's perfect. I was gonna have just like dial through what those were but it's it's really like kind like a six-step six-step prompt to prompt Chat gpt or to prompt in general the two things that stand out to me is the reasoning like when I look at that I think a lot of people are like have piece-nilled this together themselves role task

context, reasoning, output, stop conditions. The two that stand out to me is reasoning and the stop conditions. Like I think a lot of people are now starting with the role, like here's who I am or here's who I'm trying to be. Here's what I want you to do, the task. Here's the context of my industry, of my scenario, of my role.

The reasoning one is interesting though. It's another layer of detail that it needs to articulate and emphasize exactly what you're looking for. The output I think is natural. Here's what we want it to generate. But then also the stop conditions because we've all experienced that for better and for worse where the AI just starts running on its own. It has a mind of its own and it'll take you to the ends of the earth. But that's neat that you can prompt it and give it those stop conditions or tell it how to

stay reigned in rather than letting it run long, even though there's some value to that because it has, again, a mind of its own and can take you into new depths or new thought processes, but you can also reel it in as need be. So I'm really glad you shared this.

Matt Ericksen (23:27)
And that's really important in all of

this is that this, and this is a prompt where, you know, this is again, practice makes perfect. This is not something I'm still learning every day. You know, I, in the big scheme of things, I know nothing, right? But this is just what's been really effective for me. But before this, you know, it's two steps. It's, know, what do you want and what do you mean? What do you not want the prompt? What do you want the AI to do? And so just to find, Hey, I want this. I don't want this.

I want a nicely worded email that's authentic, human sounding and tone. I don't want a robot email. I don't want an overly professional email. And you'll just start, after playing with it, you'll understand, all right, what I put in, this is what I get out. You do that enough times, you get a sample size of how you interact with the system. And this is a great way to then articulate for much bigger projects, for much bigger requests, a great way to kind of segment your ask.

so that you get a really good result that actually meets what you originally intended to get.

Miriam Allred (24:28)
Yeah, and in my mind, this is like a shortcut. feel like for people like you and me that have been tinkering for years, it's like we've had to just kind of like figure this out. Like everybody, everybody that's tinkering, we've kind of like, again, figured out the best way in our own minds. But now it's like we're several years in. Open AI is like, here's how to do it. So anyone that's getting into the game now, it's like, this is literally how you do it. And you will see success instantaneously because this is how it thrives. ⁓ You also shared with me what you're kind of

defining as like this overall productivity index because people do have hesitations with AI. It's in I think a lot of people's minds, it's all about like ROI and so for you even personally in your role, you've had to think about

There's all these opportunities. There's all these workflows. There's all these tools. How do I personally prioritize where to focus my time and attention? And then once you find something, how to decide whether or not to lean in and invest in it for it to turn into something? Because I think there's just a lot of AI noise. And now it's getting hard for people to decipher where to spend their time and energy, which tools to invest in, where to let AI run, where to scale back and let the humans stay at the forefront.

Like you've been thinking a lot about this, like where do we lean in and where do we stay put? So you defined this like index. Can you explain what that is and why that's important for people to basically analyze like their AI productivity?

Matt Ericksen (25:58)
Yeah, I think that it started from.

you know, and this is kind of how I'm wired, know, ADHD mind, dyslexic mind, like I just, know, shiny objects could get me, right? I'm not gonna lie, I'm a sucker for them every time. And one of the things that you build in is, you know, as you train yourself to kind of work with that, you're wired there is what's the end result of this? Sure, it's interesting, it's curious, but does it actually align with your company's strategic objectives or does it solve for a problem? And...

That's really where that was the premise by which, you know, all right, when I got into AI, I'm like, sure, this is all super neat. I could lose lots of hours of sleep on this thing just playing with it. But where does that jump happen between just playing with something that's neat? It's a trinket. It's maybe it solves a very minute task for you. But what do you make the jump to actually scaling, deploying beyond just your internal role to everybody else? Right. And so as I've

progressed in understanding the AI tools that are at our hands and what we can actually build with them is this overall productivity index or OPI is what I call it. I put everything through this and I'd like to really look at the overall labor cost and what the cost of the AI tool that potentially could be deployed is. okay, if it cost me say X dollars to

or build this AI asset, okay? How much labor time and hours and how much I'm paying in payroll can I move from that task that the AI saw for to a higher value, higher productivity, higher ROI gain that my staff, my team otherwise couldn't access because they were.

they were limited by the manual repetition of the task that AI solved for. And I look at it as like, okay, so, all right, what's my average? Just take your average labor. If you have somebody that's say salaried, divide their salary for the year by 2080, get their hourly rate, right? Ask them how much time on average are you spending this week on this repetitive task, right? Multiply that time investment.

by their labor costs, under break down your the cost of the AI agent or the asset, figure out what that is, you know, on a weekly cadence as well too. And then, all right, there, and it's almost always a win, almost always a win, whether you have to build it yourself or you have to acquire it from a software provider. But then I'd much rather spend that say three hours of scheduling, you know, labor, instead of having to do manual note taking.

I'd rather than talk to clients. I'd rather than talk to caregivers because I know client satisfaction, caregiver satisfaction, it solves everything. If they're happy there, they're happy, you know, the business takes care of itself. So why not try and move my team and my staff to those high functioning activities that generate a way better ROI than being locked behind a computer taking notes? Paying them the same.

Miriam Allred (29:17)
Can you, yeah, super well said. And I think this is so valuable for leaders to hear this because again, a lot of people are in that tinkering playing. This is a nice to have. This is helping one person do one thing and it's cool and like let them explore and tinker. But you said it so well. At what point do we...

Does it get baked into our processes and it becomes a new standard and everyone is going through this same standard or this same operation using the same tool and function. So I think that's a really important distinction. Can you share a specific example? I think even with your usage of copilot, that has been an adoption process. so figuring out the math and the ROI of using that tool and streamlining it, can you just share some numbers to illustrate this formula?

Matt Ericksen (30:08)
Sure, yeah. the single biggest, ⁓ I would say, case study on this OPI for us was we talked to a lot of franchise owners, their teams, any given basis. have a team of four alongside me here in the operations department. And we take a lot of notes. ⁓ We do a lot of franchise calls, franchise support calls. And it's a lot of content that we go through on any given day.

It's also a lot of notes that we have to put into, say, our FranConnect, which is our operating system. So just like our franchise owners use WellSky, we use FranConnect. That helps us stay on top of our notes and our jobs and do the best we possibly can. And I asked everybody, hey, help me understand how many hours you spend. I know it's hours, not minutes, it's hours. How many hours do you spend per week? Not talking to anybody, just putting notes into FranConnect. And I found out it was coming out to about, and I have the notes down in front of me here.

about three and a half hours a week per staff member. We all work over 40 hours, but let's just say it's 40 hours. That's a significant chunk, if you look at it, of three and a half hours that I could be talking with somebody else. I could be talking with more franchise owners. I could be working with my team collaborating at a higher level. So how can I solve for that? Well, we know that Copilot has a transcription to talk about the voiceover IP, and Copilot presents that. It's an add-on service we can add into our Microsoft environment, and we did.

I tested it out first myself and I figured out okay. Well, yeah about three and half hours for me too. And it takes me about 10 or 15 minutes to document one call manually. It takes me 10 seconds now to copy and paste the transcript that has way better notes than I'll ever take from into an email. Send that email to the franchise owner so they know what our takeaways are next, action items are for the next call. I know by the way, and then put it right back in the Frank Connect and do that in a millisecond compared to what.

we were doing before that compounding, right? So across all five of us, three and a half hours each, right? With our average labor cost around $47 per hour, we were looking at $822 per week that was being spent or 17 and a half hours being spent on just notes. And we all have, we're all subject matter experts.

Why would we have the operations team spend 17 and a half hours altogether on manual note documentation? We need to do it, has to happen, we can't skip it. But this, by putting it through this overall productivity index, we were able to figure out that, all right, it costs pennies, quite literally pennies, $2.75 per person per week to have the copilot transcriptions. That's a big...

immediate ROI and that overall, that OPI is the best way to figure out, okay, can we make this thing work for three weeks? If it works, great, it will keep it. If it doesn't work in three weeks, we can't find a way for us to implement it, adopt it, absorb it into our environment. It's not worth trying to put, right now, trying to put a square peg through a round hole. We can revisit it, but we need to move on because we can't go down the rabbit hole and then reduce our performance because our goal is

support our owners the best we possibly can.

Miriam Allred (33:35)
I like too that you even just referenced that like three week timeframe,

Because it can become a distraction. It can become noise. It can become additional work if you're focused too much on these things. it's like focus on it, concentrate for three weeks, go all in on it, track all of this, and then you either keep it or you kill it. It's like there's no, we're not wasting time or energy. It's like it's either it's in or it's out based off this overall productivity index. Well, you just say kind of like the one line formula of the OPI index so people can, listening to it can understand exactly what the formula is because I think it's really powerful.

Matt Ericksen (34:08)
Yep, I'll

Miriam Allred (34:09)
for both.

Matt Ericksen (34:09)
try to articulate the best I possibly can right here. ⁓ So what you wanna do is take, know, you know, OPI is simply hours reclaimed times the average hourly cost versus the weekly, let's say in this case, weekly AI tool cost, right? So you need to know your hours spent on that task. You need to know what your labor cost is for that period, whether it's daily, weekly, monthly, whatever, maybe define the...

time to period of time, it's hours reclaimed times average hourly cost compared to weekly AI tool cost. And that's your OPI.

Miriam Allred (34:46)
Yeah, perfect.

Perfect. I'm just thinking of people listening to this. It's like, okay, what exactly is this? They can write that down so they can do it. No, you're doing great. No, is no fantastic. And you painted it in with like concepts so that people understand like what you're talking about. But I also want to just like define it so they can write that down, take it to the team and be acting on that. ⁓

Matt Ericksen (34:50)
and I can be all over the place, so no, yeah, absolutely.

And one call to action I want

to throw in here to leadership because this is key and I've made this mistake and I can probably continue to make the mistake because my brain runs just so fast. I have to slow down and really be very intentional with how I explain and introduce really innovative stuff in is as leaders. If you have direct reports that you're looking to roll out these new tools tool, you have to take a bet. You may know it. You may be a certain level of familiarity with this stuff. This is game changing stuff that has been, you know,

50 years ago they were talking about in sci-fi novels, right? It's not gonna be learned by everybody. Everybody is wired differently and you have to really think, how can I best introduce, explain and paint an illustrative picture of the value of this new tool and walk slowly with explaining it to my team? Because this OPI won't be an effective measurement if we are as leaders not effective in explaining.

how we intend for it to be worked into our work life and where we intend this thing to go. So we have to be really, really intentional and well thought out with our introductions of this if we really truly wanna measure the right ROI on this tool.

Miriam Allred (36:13)
And the word that comes to mind to me is that that multiplier word, explaining it to people as this is a multiplier and natural language. Like I guess those are the two keywords for me is like multiplying and also natural language. Like this is not some new technical skill that you're teaching people. is natural language. It's getting back to the basics. It's getting back to natural language. I love that line you said a few minutes ago, like with Google, we

have become wired to write SEO bots because that is the frame of reference that we were in. And now AI, it's actually stepping back from that mold of SEO to natural language. How do we think about things? How do we explain things? I love psychology and thinking about the way people think and getting back to this natural language way of thinking is pushing us all to be more creative, be more communicative, be more explorative with our thoughts and with our words.

I actually love this transition back to like vivid thinking and vivid communicating because that's how AI thrives. ⁓ Let's get into a couple of workflows. I love that we've like set all of this context and you've already like shared these formulas. There are a few specific things that you are doing and have done and you're gracious enough to like share some of these workflows that you've built. Let's start with the job ad grader because I think that's an easy one that anyone listening

Matt Ericksen (37:18)
Yeah.

Miriam Allred (37:40)
to this can basically like run with and go and implement in their own organization. So let's start with that workflow of like what is the job ad grader? What does it do? How did you build it? And how can people replicate this?

Matt Ericksen (37:51)
Sure,

so within, I'll use OpenEye's chat GPT as ⁓ our environment here, right? But Copilot, a lot of your LLMs are gonna have ⁓ a tool now, you can build your own agents, can build your own, whether it's agent, application, the title will vary by platform, but you're building your own app essentially within their environment. ⁓ And so this job aggregator is one of those internal, say, ⁓

GPT is what they officially call it on the OpenAI platform. It's a GPT that my internal team uses to help solve for the pain point of we had a lot of requests for, can you look at our job ad? Hey, we took your template. We have a base template, right? We're not doing one off all the time, but we have a base template. Everybody refines, but how do we get it little bit better? Job ads are both science and art. There's a science to writing it. There's an art to finessing it. Well, that can take time. How do I?

You know, my goal was how do I shorten this? And so going into the GPT creating tool, essentially, played with it and was like, all right, I need to gamify this. Nobody likes, getting critiqued and, and, having come across harsh. So I got to run in the opposite direction as best as possible. How do I do that? I need to gamify it. Let me assign a grade. Grade is easy to understand. We are all universally kind of conditioned to understand what an A minus means versus a D, a D minus, right? So, I was like, all right, checkbox. I have a, I have a, a,

a method of explaining, improvement or increase or decrease in performance of the ad based off that conditioning. And the, you know, I need to make it to ultimately save time. lot of my, themes here is, the agents I built, the tools that my team has at their fingertips is it's a safe time and to create more relationships and more conversations. This job aggregator is going to save us, it saved days. I haven't done an exact calculation with it.

like I did with the co-pilot, but we essentially take a hyperlink to our publicly posted ad and we will then paste it into this GPT. It will grade it. It will give it an A minus, a B plus, a B, whatever it may be. It will highlight what you did great. I'm a big proponent of we need to reinforce, lock up, freeze, do not touch what you did great because we need to, don't fiddle with that. Fiddle with what comes next. Here's where you can improve.

And that's what the grader does. It locks up what you did great. Don't touch it. Don't mess with it. It's good. Leave it be. And then here below it, this is where we want you to focus. And this is if you make these changes, actually the prompts tell you make these changes. You should be able to get go from B minus to B plus, right? There is no and I'll mess with there is no A plus. I built it so there's no A plus. A plus would imply perfection. There is no perfection. We don't have the black book on indeed's algorithm. So that'll never occur.

So the best we can do is an A minus and that's what we just condition our owners to understand is A minus is where we want to be at. If we can get you an A minus, we've solved pretty much as much as we possibly can. And this has helped them fundamentally maximize their organic job ads and it's helped them on a cost savings. They spent less on sponsoring their job ads to mask essentially Band-Aid what they didn't understand or have the awareness of what makes it good at. And so we've been able to not only increase time savings,

but to also decrease expenditure on our franchise owners' with sponsored ads because the organics perform better than they ever have.

Miriam Allred (41:16)
Incredible. To get slightly more prescriptive, you shared with me a couple of prompts that people can literally like cop and paste into their own chat GPT. Can you read out two of the prompts that you shared with me?

Matt Ericksen (41:28)
Yeah. let me find my notes here. So I wrote so many.

Miriam Allred (41:31)
Or I've got them

too. I could share one of them is grade this caregiver job ad for clarity, motivation, and fairness. Give me a letter grade and five prioritized edits that move the needle in home care. Like that is the level of...

clarity that we're giving the prompt. So the other one that you shared was rewrite with strong A minus. Yeah, read that one out.

Matt Ericksen (41:51)
I have it for you. So this

is kind of good talking to the prompt engineering side of things. And this is just where like, you know, it's never evolving paths. So nobody ever I think arrives at perfect prompt prompt to come up with but is you know what you want, but you have to articulate that. And that's where like if you're know, if you're an operator, wherever, wherever you're working, if you have somebody with a history degree, somebody with an English degree, somebody that knows how to write, I can guarantee you they are probably going to be your go to default.

prompt engineer in your office. You don't have to go hire them. They're just naturally going to able to articulate thought better than anybody else because of that degree that they have. And it comes down to this. you know, rewrite to a strong A- using plain English, clear pay, five tight bullets, and a 35 word intro that speaks to purpose and flexibility. So we all want that. We all think that. But how many of us can clearly articulate that I just spoke to it? That's a prompt, that's prompting engineering at its simplest terms.

Miriam Allred (42:41)
Yeah.

Matt Ericksen (42:49)
And so it's just that reconditioning your brain to say what's already in your head, but to do so clearly and concisely, that constitutes a clear high quality prompt.

Miriam Allred (43:00)
I could not agree more and I love that's why I wanted to have you read that prompt because it is like perfect like that is like a perfect prompt because like you said This is what we want. This is what we know we want but it's another thing to write a prompt that gives us what we want and And I love what you're saying too, like shout out to all like the history majors and the english majors Some people may think like oh they're all going away because of ai But they're actually the ones like kind of like thriving with ai because they are good at natural language. They're good at articulating and communicating

I love ⁓ this so much. Let's go on to the second workflow actually, which is leaning into your copilot with Microsoft. And I know that's a little bit specific, but I think people can understand like what that copilot does. You have built what you call Jean, which I love, which is after the founders of Griswold. You have named and built this copilot for internal operations with your team specifically. And I think this is

another really powerful example. So explain Jean, and the co-pilot and how you and your team are using this.

Matt Ericksen (44:05)
Yeah, absolutely. I'm a big believer in, simplifying, keeping everything. We know why I have a million different software platforms, right? And so Copilot's been our most recent kind of strategy shift where we're just trying to merge everything into a Microsoft Office suite. And I'm really happy with it. I know some people have their feelings about it, but I honestly can't say enough good things about the tech stack that we have built. with Copilot, you know, because now we're bringing that, that

generative AI functionality within that Microsoft space that we have built here. We can securely connect it amongst all our SharePoint, our OneDrive, all the areas that we have the secret sauce to what we do here, right? We have our forums, best practices, processes, everybody does. It all exists in some way, shape or form in each of our ⁓ work environments, right? But it has the ability now to be layered across all those. And what we...

One of the other pain points I keep finding, and we're fast growing, we have 40 plus new franchise owners under two years in business with us. We need to rapidly deploy and articulate very clearly what's best practices, what's the best resources for an owner that is just learning and being immersed in this world. ⁓ The faster that my franchise business coaches and I can provide that information, the more effective we can be in supporting them, but also we can reach more of them in the same amount of time.

That was the premise for I need to create a ⁓ companion tool, an asset that can help me reduce time to essentially tool acquisition or document acquisition. Because we know it's there, but you know, and I'm sure everybody can resonate with this thought is how many times have you messaged somebody in your office on Slack or on Teams with, hey, where's this form? Hey, do you have the link to this file? Now compound that, add it up. How many times over the course of a year do you do that? Those seconds and minutes add up to hours.

lost productivity, and we just gotta be 1 % better than everybody else. And so if we can be 1 % better here and I can talk to more people than you, I win. And so we put together this agent and we named it Jean because I just love the stories we're too good together. know, Jean's mission when she first started Griswold was to help as many people as possible. Well, in 2025, you know, we have an AI tool that is later across our entire ecosphere and its ability to help us help more people.

is profound. I can talk to more owners, I can talk to more other teams, I can provide more insights and more resources more accurately. And so why not call it Jean? And I thought that just it's a great way to kind of keep her legacy vision and the culture that we have here at Griswold alive and well in technology days.

Miriam Allred (46:46)
I love that so, so much. And my mind goes to this place of like, there's a lot of AI tools coming on the scene and we're almost getting to the point where they're like siloed and fragmented and it's getting like convoluted. We've gotten to this place really quickly and I actually think part of the next couple of years is we're going to see more of this where there's almost too many like point solutions and AI solutions that we're trying to piecemeal together. But the concept here with this co-pilot is you all have like Microsoft as your base. All of your tools, resources,

protocols, regulations, all of that is housed in one environment. And the copilot is the layer on top of that that can consume, scrub, analyze all of that information. And I think the application for anyone listening to this, like it may or may not be Microsoft and copilot, but it's get all of your resources, all of your internal documentation in one place, layer an AI on top of that.

And like you were just speaking to the unlocks of time and ability and capabilities of your team is multiplied profoundly because just the speed at which you can access information is insane.

Matt Ericksen (47:51)
Yeah.

And if anything, I'm from the intangibles that are hard to measure to is your team's happier is what I'm finding is that, the more that we're given to reach our full stride with these tools, but it just, it reduces stress, right? Anxiety, anxiousness, you want to help and you want to do right by the franchise owner or their teams. And, but you have this, this pressure with time that, you know, if I can help you breathe better and breathe easier and sleep better because you have time to do these things.

That's something that sometimes it's hard to measure that, but it's tangible.

Miriam Allred (48:25)
I'm really glad you added that because yeah, job satisfaction with AI. Again, I think there's a lot of like fear and hesitations and uncertainties, but I'm really glad you just painted that in of like, no, job satisfaction is going up. Productivity is going up.

maybe time off is going up. It's like all of these ⁓ more like qualitative results from this very like tangible technology that we're using. Can you share? know you've got it in your notes as well. Just a couple of prompts that you've used inside of Jean again to kind of like illustrate what you guys are doing with this tool. Can you share a couple of those?

Matt Ericksen (49:00)
Sure, so I'll give you an example because like I mentioned earlier, Copilot is we use it to help transcribe all our calls and then taking has just been so profound in helping us just do a better job of what we do and working with our franchise owners. And so we have just...

in one hour's time, 30 minutes time, you can talk about a lot, especially me. I talk fast. I can brain dump a lot. My owners roll their eyes. If anybody's listening, they know what I'm talking about ⁓ is, okay, what the heck did Matt just cover in 30 minutes? Well, prompt being like, all right, summarize this meeting for franchise owner, whoever it is, right? In 150 words with three decisions, three action items, owner of that action item plus due date.

and a bulleted risk section. What happens if we don't do this? Which is something that I learned. Really, it's a quiet killer. What if we don't take action?

What's the implications? Those can be oftentimes, those are almost always worse than trying and failing at something because it at least it mitigates it, right? And so we've always kind of leaned into how do we help everybody understand we've got to get this done. And it's not because we're just want to tell you to do something or it's not because we're assigning this to make more work. If we don't do this, it doesn't align.

we won't complete our mission of helping as many people as we can. We won't reach our annual goal. And that can really help, especially with a scheduling coordinator, a recruiter, somebody that's heads down on a weekly cadence. This helps them understand how much and how important they are to the bigger picture. At 10,000 feet, where you as a leader are operating on daily basis.

Miriam Allred (50:31)
I just love how like tangible this is. And when you define like those prompts, it's like, wow, people listening to this, it's like, take this and go apply it in your own organization. Like this is how, this is like a blueprint, like a framework, from like Matt's brain and Matt's experience, but translate that, take that, let your team listen to this and let their mind just run on like the applications for them personally. I want to share one more workflow that you've shared with me, which came out of a call with you and a

franchisee,

you guys were just like riffing and having one of your meetings, you know, just like talking through ideas and concepts and you came on this companionship exemption coach. And again, could you just tell the story of like the meeting that you had, the, the, talking points of what you guys were working through. And then you were like, wait, I know this tool that can do this thing. Like, let's just like build this thing together. Can you just like kind of tell, tell the story of how you landed on this really incredible concept?

Matt Ericksen (51:26)
Yeah, yeah,

we were, I was actually on the phone. It wasn't with the franchise owner, but it was with just a consultant because we were just, you know, like everybody right now, you know, on July 2nd, you know, the FAB came out on the workforce exemption, which materially changes, you know, our operating environment since 2013 when it really went into effect. you know, there's a lot to unpack there, not only just materially what the direct implications are, but then also the ripple effects as far as how do you operationalize this? How does this impact caregiver?

Wages, how does this impact your margin? How does this impact scheduling marketing? It touches quite literally every piece of the business. I'm like, I'm having a tough time wrapping my head around this and I live in this world. This is my job to understand this. And if I'm struggling with it, God only knows, you know, know, a scheduler, a marketer, a franchise owner, you know, my team themselves, you know, there's too much on your plate to try to decipher this. So I'm like, there's a better way. And I've been playing with

the coding in GPT-5 with the rollout of it. And it was really like that I saw, so I'm like, you know what? All right, I'm on a call, 30 minute call, like as my brain start taking off, I need to understand A, what is this consultant talking about better? So first I'm like, okay, GPT-5, hey, know, and I was sloppy. It wasn't a well-engineered prompt. was like, hey, I need, give me a ⁓ clear breakdown of the July 2nd FAB. ⁓ Help me understand this thing.

Quite literally, that was a prompt. Like, I'm not saying that's best practice. It was just, I need to fire off real quick so I can keep pace with what the consultant was sharing on the call. And it gave me something. It gave me something better than I thought it was there. I'm like, okay, great, understand this. Okay, now, and this is where I'm fiddling. I'm quite literally fidgeting. I had no idea that this was gonna work or not, but where does this thing go? I had it put together some code for me that I could paste into a platform I like to use called Replit.

It's a software development site that uses natural language processing to essentially code, right? You can also paste code in there, but too technical. I'm not going to into that technical rabbit hole. But essentially, it gave me the code and it provided essentially the framework for an app. which was the best way I can summarize it is in 30 minutes, I took a one page FAQ, a GPT scaffolded in app module.

And before the meeting ended, I had a fully functioning, know, cloud-based dashboard app that I could deploy to educate myself, my caregivers, anybody that needed to see this thing, really, know, franchise owners, executive team, at the state level, what the risk factors were and what, what you needed to be aware of. And I'll share my screen. ⁓ And this tool is, ⁓ it's been extremely helpful with just,

creating situational awareness on something that's real time, that's going to impact us quite potentially, right? You we're seeing indicators that this thing most likely will go through and we need to be prepared. And so, this is the tool that I've put together where you can change the, you know, everything from the franchise location, you change the operating state, we're gonna get interactions, it's gonna give you different risk factors, different levels, thresholds. And so I'm gamifying, I'm creating a very, you know, friendly UI user interface.

to help with us for understanding compliance guardrails. If this happens, what does scheduling look like? It has a scheduling simulator in there to help us essentially learn faster what's an optimal schedule on the other side of this reality if it occurs, right? What does pay look like? What does payor source interactions look like? What does referral marketing look like? And to derive this from a sloppily engineered prompt,

I got a really great breakdown and converted it to code. This all was built in 30 minutes from an initial, just tell me about the FAB to I have a viable asset that I can deploy to my executive team, that I can deploy to anybody that we interact with to help them very quickly understand implications and how to approach this new version of the industry if this thing goes through.

It's just about better positioning. How can we help our owner succeed? And there's many more versions of tools like this that we've put together to help our teams just understand what's happening and ever increasing speed by which this industry operates.

Miriam Allred (55:58)
When you first sent me this, my mind was blown. I was like, wait, Matt, you came up with this, built this in a matter of minutes. Just to toot both of our horns for a second. This is why you can't get this information elsewhere. This is why this podcast and people like Matt are just changing the industry because we're sharing this level of information publicly and you, are sharing this so graciously. Quick plug for all of you listening to this, Matt's sharing his screen right now and on YouTube,

can actually see what he's showing.

In real time, it's actually live, a tool built that he's using with his network. And back to the beauty of natural language, you had this call with this consultant, kind of had like these abstract concepts in your brain, put it in chat GPT, it gave you better prompt, you put this into the tool called Replit it builds this for you. Like this is not five, 10 years from now pie in the sky, we get something like this is today, you have built this tool, you are using it proactively.

preparing

for the FAB rollout. Like, it's nuts. It's just mind blowing. And I just want to thank you for sharing this first and foremost, but also hope that everyone listening to this can just like start to conceptualize the reality that we're in with AI right now. It's insane. And it's here. And it's amazing. I'm preaching to the choir, right? You are you are all over this.

Matt Ericksen (57:19)
Yep.

Yeah, but it's getting the word out on it and it's not, have a history degree. like, quite literally, I am the antithesis to who should be on paper qualified to do this and I'm finding ways to win with it and support our network and our owners and I use it as a call to action. was like, quite literally, if I can do it, and I didn't have a great GPA either, but I survived college and high school, and if I can do it and I can figure this out, then quite literally, just put some time aside in the evenings and just.

fiddle with it, tinker with it, it's gonna present ideas that you never thought of. That's the glory of it, that's the value of it, to where you're like, it's gonna get your creative juices flowing and you're gonna find ways to get excited, to help just find ways to support your teams to success. ⁓ It's only gonna make everybody better, in my thought. It's all mindset and how we open to change and learning a new thing.

Miriam Allred (58:15)
That makes two of us B / C students, but here we are today just like rocking it in home care and changing the game with AI. You've shared so much good information today, Matt, and I hope people are just, I love these conversations. They just get people to think differently, like step outside of their current thinking and let their mind run on what's possible with AI. Just a couple closing questions. One is,

Where do you think we're gonna get to? Like, let your mind go to like 2030, like five years from now, like look how far we've come in three years. Where do you get excited about us being in maybe five years from now?

Matt Ericksen (58:54)
This is extremely dangerous question because I could talk all day,

every day on this one. ⁓ But where I really see it, and I think specific to the industry we're in with home care and so forth is, right now we're in a stage of initial understanding, but more so on the operational base side of things. ⁓ We're just emerging with getting involved with the generative AI for visuals. It's still on its way there, but it's not perfect. So operationally,

⁓ It's having its biggest impact there. It will have a very big impact in the near future because I can talk about compounding It's only the speed. It's not gonna be we're not gonna ⁓ We're not heading into the future at a constant rate of speed. We are heading to the future at an ever-increasing speed So if we're running at 60 miles an hour today, we're running at 120 miles an hour next year and in two years time What I think we're gonna see is we're gonna see the operational workflows best practices with the adoption of AI

be an afterthought if you're not doing it, behind the eight ball two years time. What we're gonna see then is really the effective use of that technology, best practices, being assimilated into now the healthcare setting. How are we taking PHI information securely? How are we taking the clinical world and the diagnosis codes, the ICD-10s, taking all of the clinical care notes? And how are we in a more broad stroke?

Does that world intersect with non-medical home care? How does the AI bring together words that have historically been very far apart as far as collaboration only recently came together? How do, you know, it's the clinical element information that's now available in their world and they're using it, but it's making that jump to where they continue, much like how the guide program through Medicare is functioning, where we're trying to bring together a lot of different stakeholders. know, AI, see that bringing together worlds that have historically been siloed.

to better focus ⁓ on care outcomes and to get there faster in real time. I'm really excited about is the actual the hardware application. We're talking software on this conversation. Really there's a whole hardware element of this as well too. And it's not a robot to replace a caregiver, ⁓ but it's a how do the hardwares blend in with medication reminders, medication disbursements, remote monitoring systems.

How do all these assets kind of come together into that environment around on a patient focused environment? That's where AI is really, I think, going to find its pace is it just it's bringing people together. So as much as we've we're used to the Terminator movies, where, you know, there's this other side there where this has a very real impact to get people off their phones to have, and not be siloed in their own little universe is it's going to pull together a clinician, a caregiver, a counselor.

know, behavioral health, and we're all gonna be able to talk, so much more effectively with each other and better understand each other. I think we're gonna better outcomes just from knowing exactly what you're thinking and even if you're a thousand miles away.

Miriam Allred (1:02:02)
You know we love this stuff when that just like gave me chills. It's like, it's so cool to think about where we're gonna be. And I love your take on this, like the medical field and the home care fields like coming together in a way that we've never been able to before. You know, we've been working so hard on that relationship and on rewriting that narrative and the continuum, but it's like, wow, AI has the ability to bring us together in a way that we've just never known was possible. And that is really exciting.

Matt Ericksen (1:02:29)
And

one thing I gotta add to it, because I forgot to mention it earlier, but it's the, what's, helped greatly on our side too with that, right? Is, you know, speaking up, right? So we know in the home care space, we're used to personal companion care, know, lingo conversation, right? But there's a whole, we're coming from the home health side, pediatric home health. There's a whole different language, terminology, right? And the AI tools, you

they can help you speak that language faster. You don't have to be a clinician to speak like a clinician now, right? You can find out that you can prompt the AI to, hey, I really want to impress this director of nursing at this post-acute rehab facility. And asking it, tell me what I need to know, what do need to say for my position? And it's going to help me speak that DON's language. First meeting, first impression.

create a higher impact, it's gonna help me do my job better and reach my end goal and break down historical biases or pre-existing conditions that were there. And it's gonna help as far as tone, how we interact and communicate with each other. It's gonna help us, I think, just be better offline as much as it helps us be better online.

Miriam Allred (1:03:45)
well said, gets me so excited and being a part of it is just like a pleasure. Like we are going to be a part of shaping that future, which is incredible. Last thing, Matt, because you thought about this and I want to include it as well. We've got hundreds of owners listening to this right now and what's one AI experiment that you would challenge all of them to do this week in their business? Like get specific, get tactical. What's one thing that they should go out and do this week?

Matt Ericksen (1:04:13)
Absolutely. So I had to think about this one because there's a lot of different ways we could go with putting a prompt in. But let's start with what can you affect today? What's been, and most importantly, what's been bothering your business, bothering you? Is it a hurdle, roadblock? Is it a choke point? What is the issue that's holding you back from the outcome that you desire the most? Tell that to the agent. Tell that to the chatGPT, Claude.

know, Gemini, Copilot, whichever one that you have the easiest access to and explain your position. Follow the prompt guide that I shared earlier, you know, break it down as best you can there. It doesn't have to be perfect because it'll it'll logic will pick it up and ask it, hey, this is my choke point. This is what's holding me back. I need to get to say a thousand billable hours, by by end of the quarter, right? Provide me here's prompt. Provide me a 30, 60, 90 day

Strategy plan to help me overcome insert primary choke point or primary hurdle insert and then by this date insert the date that you want to achieve your objective by and It will provide you most likely provide you a high level one if that's all you provide the more information that you provide from a context standpoint from a reasoning standpoint from a task standpoint and you just beef this thing up that That basic prompt can yield for you a very actual plan

that you can implement tomorrow, today, Monday, and that will, oftentimes we struggle in the weekly cadence that we exist in within this industry. Most of us are on weekly billing cycles, is breaking out of that hamster wheel effect. Every week, it's the same thing every Monday and Tuesday. Billing and payroll, after that, it's moving on to how are we looking going into the weekend? Mid-week orientations. We are creatures of habit in this industry, and that's why it's making it so hard for us to

force multiplier, our pursuit of our overall goals. This can do that for you.

Miriam Allred (1:06:19)
That's why I wanted to include that because the reasoning is so powerful. You could have put together a prompt for billing, for recruiting, for payroll. You could have, you spend so much time in this and I wanted to push you to answer this question because you work with hundreds of owners. This is the most important thing is getting out of that hamster wheel, stepping back and thinking about strategy and breaking down barriers. That's what leaders are struggling with and this is a AI prompt to help you break down that wall in your own business. Matt, you're phenomenal.

I think the world of you, you delivered. I pushed you to come up with some of this and to map some of this out, but I've been so impressed with how you've articulated all these concepts and ideas. I have a feeling people are gonna want to reach out to you. Obviously you're busy and your time is precious, but what would be the best way for people to communicate with you in a way that's conducive to your schedule?

Matt Ericksen (1:07:07)
Sure, so

work email at matt@griswoldcare.com and I'm very active on LinkedIn, so just search Matt Erickson, again, Vice President of Sales and Operations here at Griswold and based out of Bluebell and ⁓ if anybody's whether you're local, you wanna connect, I always love just learning from everybody and connecting with as many people as possible. So excited to be here, thank you for this opportunity. It's been an absolute guilty pleasure just to talk about the things that I love working on and. ⁓

I appreciate the ability to share everything that I love doing here.

Miriam Allred (1:07:40)
The feeling is mutual. Thanks a bunch for joining me in the lab, Matt. Until next time.