Boost your brand with The Business of Pharmacy Podcast™ – Click here to advertise
March 25, 2024

Automating Compliance in Pharmacy | Sumeet Singh, LighthouseAI

Automating Compliance in Pharmacy | Sumeet Singh, LighthouseAI
The player is loading ...
The Business of Pharmacy™

Sponsored by Parcel Health: In this episode of The Business of Pharmacy Podcast™, host Mike Koelzer welcomes Sumeet Singh, founder and CEO of Lighthouse AI, to discuss the revolutionary approach his company is taking towards automating pharmacy compliance. Singh delves into how Lighthouse AI leverages artificial intelligence to simplify the complex and ever-changing landscape of pharmacy regulations, ensuring pharmacies can efficiently navigate compliance requirements across different jurisdictions. This candid conversation sheds light on the technology driving the future of pharmacy compliance, promising to transform the traditional, labor-intensive processes into an automated, efficient, and error-free system.

https://lighthouseai.com/

www.parcelhealth.co/

https://www.bizofpharmpod.com/

Thank you for tuning in to The Business of Pharmacy Podcast™. If you found this episode informative, don't forget to subscribe for more in-depth conversations with pharmacy business leaders every Monday. For additional resources and updates, visit www.bizofpharmpod.com. Together, let's navigate the ever-evolving world of pharmacy business.

Transcript

This transcript was generated automatically. Its accuracy may vary.

[00:00:31] Mike Koelzer, Host: Sumeet, for those that haven't come across you online, introduce yourself and tell our listeners what we're talking about today.

[00:00:39] Sumeet Singh: My name's Sumeet Singh. I'm the founder and CEO of LighthouseAI. AI is, focused on automating and enabling the compliance journey for pharma and pharmacy. 

[00:00:52] Mike Koelzer, Host: Sumeet, whenever I talk to some of my kids who have either gone on in [00:01:00] science or those that haven't, they always talk about some classes being the weeded out classes . oh, physics at Michigan Tech, that's the one that separates the men from the boys and biochem here, that's the one that separates these fields. When I think about some of the mandates of these permits and licenses and all this kind of stuff, it almost seems like even if you had nefarious. Purposes, like you wanted to make some DEA fake thing up or something. You'd get so sick of all the paperwork and the forms and all that kind of stuff that you would just say, ah, screw it.

And then the DEA sitting there and they get all your stuff in and they just throw it out. They say, oh, that person's good because they went through it all. We're just gonna throw it out. And the ones that didn't ever send it in are the ones that are the bad guys. I'm joking, of course. But do you ever see that they make [00:02:00] things more intricate just to weeded things out?

[00:02:06] Sumeet Singh: Perhaps that could be a component of it. I really think what it is it's more about the execution of such a thing, 

we've even been engaged by , over 700 customers to date, in the last eight, eight plus years. And what we used to do, or our approach used to be, and this is the approach that everybody else has, is, you have a person and that person is in charge of understanding all the unique aspects of the company.

20 years ago, that was okay.

10 years ago it was getting almost impossible, and today it's definitely, almost virtually impossible for somebody to keep up with all these changes on a, 

on an ongoing basis. So what we do is whether it's a new company or it's, an existing Fortune 500 company that's been around for, decades or even centuries in one case, what these companies are looking for is they wanna [00:03:00] automate those, figuring out what they need to do, and then making sure that it gets done 

because compliance is really only two things. It's what do I need to do? If you're talking about drug pharmacy compliance, it's okay, what are the registrations I need to get? What's the credentialing I need to get, what PBMs do I need to work with? What, far percent's charge licenses.

Do I need to get, if you're talking about, gee, I don't know like a light bulb, and you're going 

through under UL or certification or something, it's like, what are the max wattage I need to do? And, how much stress does it need to take 

in and blah, blah, blah. ISO compliance, HIPAA compliance, what do I need to do is the first step.

And then second step is just making sure that it gets done, having policies in place and logs in place and and stuff like that. And 

so what we did is Lighthouse AI Intelligence automated the question of what do I need to do? And so that's precisely something that we help these companies with.

So if it's a pharmacy looking to do business in all [00:04:00] 50 states, instead of manually trying to figure out what this pharmacy needs to do business. On a state by state basis, and what is the process and how much is it gonna cost, and what forms do I need to fill out and can you help me fill 'em out?

We automate all of that, and so that's the company. And so I guess that's the very long answer. The short answer was yes, Mike that's exactly what we help with is all these owner's requirements that are confusing and so jurisdictional specific, we just automate it with technology.

[00:04:33] Mike Koelzer, Host: So Sumit, we hear a lot of AI now, and from my understanding, AI has gone back a long time depending on how you define it, now we've all seen AI with language models and so on, and so that's where we normal humans, it's kind of come into our face more. What does it mean when you have AI in your company?

 How have you embraced that [00:05:00] and defined that?

[00:05:01] Sumeet Singh: something really important to understand is to us, all technology is kind of like a toolbox, there's problems that you need to solve and then there's a toolbox in front of you. How are you gonna solve this 

problem, and so when there's, so that's one thing to remember.

The other thing to remember is there's different types of ai. The oldest example of AI is way back in the 1960s. They used to work through like organic samples and figure out, the organic chemistry behind specific samples using a really quick, what they call an expert system, 

You fast forward a decade or two and all of a sudden IBM has OCR, which is

Optical

character 

recognition. 

[00:05:49] Mike Koelzer, Host: I used to try

to use That instead of retyping stuff

[00:05:52] Sumeet Singh: Yeah, exactly. , it reads PDF documents or whatever, the iPhones have that capability now, and 

it's [00:06:00] awesome.

Like you can copy text off of a 

picture. 

Then you have this other thing called NLP Natural Language Processing. So this webinar, you offered me a transcript that uses NLP technology, 

something that's been trained, a model that's been trained

to learn to hear specific words, 

maybe even accommodate any accents, 

California accent I have or 

been Midwest accent that you might have.

And now. What we have is the advent with the transformer architecture, what we call generative ai or a neural network or a 

deep learning network. and so without getting too into the weeds, what I see in front of me is this toolbox, and I see this big problem called compliance.

Compliance. Like I said, we've kind of broke it up into two things. What do I need to do? And then help me get it done.

And so the first problem that I'm trying to solve, what do I need to do? How do I solve that problem? [00:07:00] Mike, I know we can't hallucinate, we can't get this stuff wrong.

People's businesses are at stake. There's a lot of dollars on the line. There's reputation on the line, there's contractual compliance on the line. We cannot get it wrong. And so the. Tool that I used, that I picked for this first piece of the puzzle out of my entire technology toolbox was what we call an expert system.

So a 

lot like TurboTax. TurboTax, it doesn't matter if you're in Michigan or in California, it doesn't 

matter if you have kids or you don't, doesn't matter if you have own a house or you 

don't, et cetera, and so on. You put in your unique attributes, you press a button, and then on a county, state, federal level,

Hey Mike, this is what you need to do.

This is what you owe. This is the forms you need to fill out, and for an extra fee, I'll fill it out for you and send it 

in for you too. 

It's the full 

service. That's the problem that TurboTax solved. The problem that we're trying to solve is very similar. I'm a drug [00:08:00] company or I'm a pharmacy. I want to ship to patients across the us What do I need to do?

Our per system will consume those very specific attributes. It'll ask you, very specific things like, are, do you take possession of the drug product or are you just kind of like a prescription clearing house. Are you doing controlled substances? If it's a controlled substances, schedule two, schedule three, schedule four, schedule five.

'cause each one of those has different requirements. If you're a schedule two, you need a special permit. In Florida with 

the Board of Pharmacy, if you're a Schedule five, you don't need the opioid permit, 

 Prescription drug, you need a license. In some states you don't need a license if it's a OTC drug and you're just shipping a bottle of 

or whatever. And and so our system will tell any pharmacy, any, manufacturer, Hey, this is what you need to do to be compliant in all 50 states. 

 That's the core of our [00:09:00] product, the expert system.

The second piece of it is, help me get it done 

To get it done. There's applications that need to be filled out using company information, using like a pharmacy information.

I wanna license in California. Okay. What is your facility, what is the corporation name? What is the corporation mailing address? Suite number, owner name, owner Social security 

number, which, 

pimm data, personal information 

data is very tricky to, for us and something that we spend a lot of time 

on securing. There's documents that need to be included with the application. It could be a high school transcript, it could be a background 

check result, which is very confidential.

It could be a lease, it could be, confidential or not confidential.

And so the next part of it is grabbing all the applications from.

The worldwide web

pulling into our system and then getting 'em filled out. And for that we use different technologies a little bit, OCR, a little bit [00:10:00] generative ai, a little bit machine learning, a little bit. Just good old logic step functions. 

If once this, then gather these documents, pull it all in and you can get it ready to send out.

So I think that's the biggest piece. Lighthouse ai, it sounds kind of. Hype forward, and we were able to raise venture capital 

money and stuff of that. But the problem that we're trying to solve is the problem that I've been working on for 14 years, I viscerally felt the pain.

We scaled the services company for five years. Before that I was filling out applications by hand as an intern all the way for the next five years. Our first customers that we sign on for the predecessor company of Lighthouse AI called Five Rivers rx. I was filling out applications for customers and clients across the us.

I was licking envelope, I was

taking 

Stamp and definitely I was 

working outta a bedroom in Minnesota is where I got to start. 

And then, 

I had a mailbox. And then and then slowly in 2017 we graduated to an office and then we 

graduated to another office. And then we [00:11:00] rent Ramon.

Now we have, 60 employees.

And so that's the genesis of, and so like. aI toolbox is technology toolbox, we use it all day long. Some of the other technologies that we're using right now is cloud computing, everything is being recorded. It's gonna go into the cloud, 

it's gonna go to the transcript.

You don't even notice it 

because somebody came and decided, hey not to make this too public, but 

squad cast fm, I'm gonna go and solve this problem for people like Mike. I'm gonna make 

this delightful for Mike by automating it and using all these different technologies, cloud computing and OCR and NLP, 

I'm gonna automatically create transcripts.

I'm gonna reduce the microphone feedback with specific audio, 

Software. I don't know the mechanics of 

it, but, and so that's we're like one of the first companies as, as far as I can 

see in pharmacy compliance that decided, hey, this is a really sucky, painful. 

[00:12:00] Complicated problem.

Let me solve this with technology 'cause nobody's ever done it before. 

And so that's, that was really Genesis. 

[00:12:08] Mike Koelzer, Host: The product you're offering physically the product, is that a SaaS program that people are getting? It's basically a computer program real similar, like you said to TurboTax or something.

[00:12:20] Sumeet Singh: Exactly. Yeah, it's app dot lighthouse ai.com. You get a login companies are able to configure their facilities, so if it's a if it's a pharmacy based in Kentucky, I just keep picking that 

example. I dunno why, It's a pharmacy in Kentucky it's gonna take about probably 15 minutes to configure the specific facility.

Hey, these are the products I handle, these are the activities. I do shared pharmacy services. I don't, I'm a 

central fill site. I'm a. Dispensing non dispensing pharmacy. I take possession of the product. I repack, I relabel, I blah, blah, blah, blah, blah, 

 I compound, I sterile compound, I non-sterile compound.

You put in all these a hundred plus compliance attributes. [00:13:00] It takes, it, it sounds heavy, but you know, our customers know exactly what they're doing in their pharmacies 

and and so forth. And then they press a button and then they're told within 90 seconds, Hey, this is, these are all the things I need to do.

In the old days, if somebody wanted to figure that out, you would go to a law firm or you would go to a service 

provider like us. Who would charge you anywhere from 8,000 to $30,000 and it would take two to four weeks. We're able to configure your facility in about 15 minutes, 

15 to 30 minutes, say, and it takes about 90 seconds of.

Computer compute time, sometimes less, 

Never more than 90 seconds to deliver, Hey, these are all the things you need to do in all these different states.

[00:13:43] Mike Koelzer, Host: How do you price that then? Is it like a subscription or is it a big upfront 'cause it sounds like a lot of that value is right upfront.

[00:13:52] Sumeet Singh: yeah. Not quite. So it, there's definitely, a little bit of a upfront onboarding and stuff of that. Of 

course. But it's absolutely a [00:14:00] subscription 

[00:14:02] Mike Koelzer, Host: because it follows all this stuff continually,

[00:14:07] Sumeet Singh: Exactly, so

For two reasons. For actually there's three parts to this and I'll go through each one. 

The first part of it is the companies themselves are changing every single day. When we bring on 

an enterprise customer, like a like a Fortune 500 kind of company, and they have nine operating companies and 26 operating business locations, facilities, there's change of ownership, change of DR change, of pick, change of product portfolio.

They bought this company, they sold that company, 

they bought this asset, they sold this asset, things are constantly changing. There's a event driven compliance like those as well as, just other kind of other kind of changes on an ongoing basis. So 

companies themselves are changing 

and they have to adapt to the regulatory framework accordingly.

The second part of it is it's a moving target. The regulatory framework is always changing. You have stuff like new drug price transparency requirements for pharma. Starting, there's been 23 [00:15:00] past in only the last three or four years. State licenses are always changing. The opioid epidemic has been 

especially onerous on pharmacies and, their wholesaler vendors and their manufacturer vendors and so forth.

So the regulation, the regulatory framework is always changing and we keep up with that on an ongoing basis, 

and so when we think about value, this new concept, we invented this new concept of 24 7, 365 surveillance, 

there's this thing in healthcare called a never event. A never event, for example, is like a surgery, a surgeon amputating the wrong arm.

That should 

never happen. In our experience, a company should never forget about a license. A company should never forget about a report they needed to submit. A company should never get a fine or a penalty or be unable to ship to their patients because they were not compliant, and so with this, with technology, for the first time ever, there's this 24 7, [00:16:00] 365 component.

We're always on. You can 

reconfigure, Hey, I'm gonna buy a. This pharmacy says, Hey, I'm gonna go expand into controlled substances tomorrow. Or,

I've been hearing a lot about non-sterile compounding. What's it gonna take for me to add this kind of line of business, to 

do some of the dermatology terminological things.

Oh, I need this specific license in here and here. I can do that. It's gonna cost me an extra X amount. 

Let me go ahead and do that. 

So when we talk about value 24 7 365, there, there's definitely the risk mitigation of it, which is 

above and beyond any cost efficiencies, but the time to market and the strategic insight can be especially valuable for pharmacies, 

as they think about how am I gonna expand my business? How am I gonna change my business? And so forth. Should I open a new location? Should I do 

X, Y, Z? So there's that part of the value that, that we're, that we talk about. The risk mitigation. Risk avoidance is obvious, you're 

in compliance more.

And so that 

part of it's less, [00:17:00] pharmacy spends so long getting additional patients, the cost of getting a patient is super high. Maybe maintaining them is not so high. 

But if you're unable to ship their prescription because Alabama or California started requiring a pharmacist in charge 

license and you don't have that and you're unable 

to ship to that patient, you have to go refer them to another pharmacy.

What's the likelihood they're gonna come back? Not very high, 

so all that investment you made in getting this patient is all gone, and so when we think about price, we think about subscription of course, but we also think about If you were in a position to never be not compliant again, 

what is that worth to you? It becomes a small tiny fraction of 

total revenue instead of, Hey, it costs me a couple hundred thousand dollars to maintain this compliance officer and this compliance 

team to to stay up on all this stuff. Can you automate that for me? 

[00:17:55] Mike Koelzer, Host: I've been scanning all my personal documents for [00:18:00] 20 years now, 20, 21 years now. And my dad always had boxes of stuff all over papers when he ran the pharmacy, and that's just how his mind worked. But I'm like, I'm gonna do that differently. So my point is I'm no schmuck when it comes to being organized that's allowed me to do other things like this, have a little bit more freedom in my mind. Every year though, Sumeet, I say when my taxes come. And you've got the quarterly estimates. Every year I say, I'm gonna keep this front and center on Google Drive, and I'm gonna do this, Michigan sucks. I think a lot of states probably suck well, they're just states, they're just government run things. their websites suck. But 

I always say, I'm gonna have the city, the state, and federal right there to click on in my folders and I'm gonna do this. Sure enough though, every quarter comes by and I'm like, how do [00:19:00] I open this damn form from the accountant?

And where's the thing for the state of Michigan? And now they're asking me to do this. Your intentions are good, but it's hard to keep up on stuff.

[00:19:09] Sumeet Singh: It's hard to keep up on stuff as the regulatory framework is changing 

[00:19:13] Mike Koelzer, Host: It's hard for the average pharmacist that's not their job to

be looking at this stuff daily. And you think you're gonna get it. But a quarter goes by, or a year goes by, or three years goes by.

It's hard when it all adds 

up.

[00:19:30] Sumeet Singh: For sure. and I think that the world is changing fast, right Mike? I think where we are now is this kind of interesting juncture where all these new technologies are coming into place 40 years ago, or even 20 years ago, when you're even really, 10 years ago when you're scanning something 15 years ago, when you're scanning something, it's ending up on a computer hard drive.

Now, in the last 10, 15 years, cloud storage has become a thing and you're able 

to put all these documents in the cloud in a Google drive 

Or Dropbox or something [00:20:00] like that.

And I think, these new notions of AI and stuff like that are just going to further, it's just gonna further automate and make things easier, 

And I won't get into it too much, but we are leveraging some technology now and even more so in the future to make that an easier process. Today we have about eight full-time employees completely dedicated to keeping up with all these changes, 

we have attorney, we have, a former regulator, the former executive director of the Utah Board of Pharmacy is on staff.

We have former, industry veterans, people that spent, 15 years at Fortune 500 

manufacturers 10 years at one of the largest pharmacies in the us. We have these SMEs, we have these compliance experts 

to stay on top of all this for us. And all we're doing, Mike is we're just feeding the system.

We're just feeding the technology. So there's, between [00:21:00] our customers and us, there's this piece of technology. That piece of technology, a, we have a all-star team of eight people,

They're wholly dedicated to that, and then we have the technology. And so the knowledge and the information from these eight people, they're gathering everything.

We have paid resources, we have free resources. We're attending board meetings. We're doing 

everything right. We capture all this information with a dedicated staff of, this all-star team, 

the best people I could find.

It gets pushed into the technology and then the technology distribute, distributes that data, 

that, 

Information, that guidance specific to the business model, product, portfolio, and jurisdictions, the a hundred plus 

compliance attributes, it's specific to that. And so it's definitely hard, but it's the concept of leverage we're 

able to do. With these eight [00:22:00] people, we're able to do what every company has historically had to do for themselves.

Instead of having eight people sitting, remotely working on this lighthouse product and feeding the 

lighthouse product, before Lighthouse AI existed,

every single one of these pharmacies had their own compliance people trying to stay on top of it. For an independent pharmacy, is an independent pharmacy attending every single board meeting of, 

their, of, say they're based in Michigan.

Are they attending every Michigan Board of Pharmacy meeting? 

You get a 50 50 state mail order pharmacy. Is that pharmacist in charge really attending all 50, 50, 50 State Board of pharmacy meetings?

No way. It's not possible. 

It's not physically humanly possible. And so by putting this piece of technology in the middle and creating leverage, it's basically 

like this, it, instead of a one-to-one relationship.

We created a many to many relationship, 

 And so that's at the crux of it. That's one of the [00:23:00] big value adds, value creation. When I talk about creating value, how do you create value? You create leverage, you create efficiencies, you create blah, blah, blah. 

That's one of the critical components is finding the technology to solve a problem in such a way that you know, you're creating more value.

24 7, 365, surveillance, 

risk mitigation, strategic insight, faster and quicker are all components of creating and delivering value.

[00:23:31] Mike Koelzer, Host: Yeah, it's very much on trend Sumit, because I've used this example many times on this show, but we'd have our accountant and I'd have to basically try to teach them about the pharmacy and they didn't understand it. What kind of person understands that you're selling most of your stuff for negative profit, things like

that.

But now I've gone to something bigger and these people are with, whatever, a hundred pharmacies and they bring advice to me. Same with [00:24:00] law firms that are more nationwide specific to pharmacists. Same

as this, because I tell you. They always say people in business, and of course, to teach people to do stuff, you don't have to do it yourself as the owner and things like that. But when some of these forms come in, I can hire someone to do it, but all that means is basically they sit right next to me and ask me a question on everything that happens. And you're like Mike, you have to train that. It's like, how do you train someone to tell 'em what your history was and how names changed at the pharmacy in 1945?

And, have you ever been, sued? I mean, There's a lot of stuff , it's a pain in the ass, those forms and the licensing, all that. So I can see the value in that Sumi tell me how with a company like yours, tell me [00:25:00] how you feel confident that you're not gonna lose all this stuff.

And I'm sure you've got AWS and you've got the Google and it's inside of a three mile island, one mile deep kind of thing. Is that ever a concern for a company like yours? I don't mean a concern because I

know you've got it settled, but

What kind of thoughts have to go into that?

Because that seems like with one switch you could wipe out a company, basically.

[00:25:28] Sumeet Singh: Yeah. Yeah. A lot of thought, a lot of 

resources We actually achieved ISO certification before we even ever even closed it's been a core aspect of our focus since the very beginning.

 I kind of mentioned earlier too not only is disaster recovery or data recovery, whatever, a critical component, but also just straight up security, 'cause we have social security numbers, we have background check information, 

[00:25:53] Sumeet Singh: we have incredibly personal information, 

 At our fingertips.

So the answer is twofold. The first part [00:26:00] of it, as far as disaster recovery goes and whatnot. Yeah. Technology is. Technology can accomplish anything. And technology is only as good as you code it, as you 

make it, and so we just simply follow all of the best practices to the

enterprise level, 

cross region duplication, for example, like if we have AWS and in two different regions, if 

one region went down and, suffered some crazy catastrophic loss,

We're running both systems at the same time.

One wouldn't even know that one went down because within seconds it would just flip over.

it's like that, and then we have, backups, we have six months live bas live backups in place. And then we have two years of backups in cold storage, where they're not easily accessible 

and stuff like that. So there's a lot of thought time that we.

Spend on this exact problem. And look, we have enterprise customers, we have Fortune 500 kind of 

customers. Publicly traded companies are not [00:27:00] willing to give you your data 

their data unless you have what it takes. 

And they do the due diligence. We've had due diligence audits, quality audits, security audits.

We do penetration testing both internally and third party to make sure that the security is there. There, there's a lot of TLC that, that goes into that whole function. And

that really is, it's just a nod to our technology team. So our technology, our CTO is actually a former CTO of vice president of engineering at Citrix.

He was there when it grew from 400 million to 1.2 billion in revenue, not market cap in 

[00:27:37] Mike Koelzer, Host: Wow. 

 I was an early Citrix guy years ago. What was it 

I forget the name of it, but it was like a, Google remote kind of thing. Desktop sharing

That was years ago.

[00:27:47] Sumeet Singh: Yeah, I'm not totally familiar with that, but that might've been before my time a little bit. But yeah, he was the CTO of, that 

from, in that time period. And so when he came on board our technology team has [00:28:00] grown to about 25 people in the last two years.

He joined almost exactly two years ago. We're coming up on his two year anniversary. So it's just a nod to him. He knows what the enterprise technology requirements are. He's. Accommodated them before he stayed on top of what it is, and he knows the best practices and so forth. We have SOPs, we have policies, and then we have the underlying technical architecture just to ensure that stuff 

is secure and

[00:28:27] Mike Koelzer, Host: almost seems like something like maybe you already mentioned it, there's some kind of national stamp of approval for a company that you've got a certain level of this

[00:28:37] Sumeet Singh: there is, yeah. And that's iso 27,001 certification. Yeah, so we achieved that in, I wanna say May of 2023. We released our product in April. We got the certification in May, and then we landed our first. Fortune 500 customer in June. 

So it was a very rapid [00:29:00] succession because we knew that enterprise companies were gonna be jumping on board quickly.

'cause out of all of our types of customers, that's the type of customer that, we solved the most pain for. 

And so, that's what that looked like.

[00:29:15] Mike Koelzer, Host: you mentioned earlier about uh, what do they call the hallucinations of ai? And

gotta be careful. I asked ai?

 

[00:29:24] Mike Koelzer, Host: for the history of my pharmacy or something, and I never knew it, but apparently I have a Grandpa David who had, you I'm just kidding.

I

[00:29:32] Sumeet Singh: Oh no.

[00:29:33] Mike Koelzer, Host: brought all this stuff out and because AI will say sorry, we can't do that because that's, uh, we don't,

delve into that. You have to talk to your doctor, that kind of thing. But if it's not something like that, they'll just make it up.

[00:29:47] Sumeet Singh: That tool specifically is called generative ai, or, large language model. Some 

 It's a type of neural network or deep 

learning 

or something like 

That That is just one component [00:30:00] over here, the tool that we used for the intelligence, the crux of our product is all circled around an expert system, which is, TurboTax, which is the 1960 that, 

so 

now what I would say is though, as a part of the toolbox and what problem we're trying to solve, there's absolutely a play for generative ai, 

For something like a chat GPT or large language model where we can leverage our proprietary data set, we've took all of the laws out there and we created out of it software code, software logic software databases, 

so we structured in VC terms, what they would like to hear is or a buzzword, I guess is we structured unstructured data, And, and that's something that, that generative AI is really good at. One thing is you always need a human in the loop, which means a human is always checking to [00:31:00] double check, they're always 

double checking. In the pharmacy workflow, this is very common P one, P two, 

there are states that allow the first pharmacist review of a pill to be electronic 

with AI or whatever you want to call it.

And then the second has to be a live human trained 

Live comma, human, comma 

trained pharmacist. 

And so that concept kind of still rings true is you have to be responsible with this new technology. It's with great power comes great responsibility. AI is incredibly powerful.

You just have to be responsible with it. You have to layer other types of technology. You have to specify which data sets you wanna work with. And ultimately, a lot of times you have to have a human in the loop. A human in the loop is another very common term these days as it relates to generative ai.

Just making sure that there's a trained human being involved In the In the process to double check and verify 

because [00:32:00] again, we take compliance very seriously. It's, it's what we for,

 For a living, is the best way to put it. 

I tell the team, there's a zero margin of error

pulling that string a little bit.

Something like Tesla, like full driving mode.

Tesla has to be or whoever is Waymo, Tesla or cruise

 division, that's doing auto autopilot, auto driving.

They have to not only be as good as human drivers, they have to be exponentially better. 

Human is getting into into a car crash, one out of 2000 rides.

Computers have to do it one outta 10,000 because it's new technology for adoption. That's the threshold. That's probably the ROI that, that humans have set. Okay, it has to be five x better for us to be willing to adopt this new technology. And that's something totally different.

But, definitely part of 

  1.  

[00:32:49] Mike Koelzer, Host: I was listening to Bill Gates talking about nuclear energy and it's. It's gonna be the way to go. And they said that it was getting [00:33:00] real close because they've got all kinds of, I already mentioned Three Mile Island, but they've got all kinds of new stuff that's very protective.

Then I think it was, I don't know, half dozen years ago or whatever, where Japan had some accidents or something and it set everything back and that was just the optics of that, of people remembering that and bringing up the past and stuff. So you're right

on those car It's gotta be a lot better for people to even. Compare it. Sumeet,

When you were talking about that, I was thinking of Google has finally come out with one called Google Lmm, and I had been waiting for it for a while. Because what that'll do is it'll take your own notes. Let's say you've got a lot of manuals in your own business, or you've got policy rules and things like that. And that'll take all that. And then that will be your source for the information. It's your own stuff versus going out to the world to do the [00:34:00] language model. It'll just take what you have internally. It kind of sounds like you might do some of that as kind of a protection. I don't mean with Google Lmm, but you've gotta be careful where your model's picking stuff up.

[00:34:14] Sumeet Singh: Yeah, for sure. And yeah, chat, GPT came out I think they call it chat, GPT Enterprise 

I don't fully remember. 

the custom GPTs and stuff like that, the concept that, yeah, you train models. These large language models on your own data? First they're trained externally, 

so you take the current instance of, barred or chat, GPT, whatever, and then you close it off and says, okay, now that you've been trained and you know how to operate and 

You understand syntax and you understand if I 

ask you a question, you answer 

If I ask for a list, what that means.

And if I say, I needed 200 words and five bullet points, you know 

how to do all that. And and then you take that and then you using the Tar Transformer architecture, which is, the big technology breakthrough,

then you focus only on your data. And so [00:35:00] I think that's definitely a part of it.

It all just comes back down to what is the problem you're trying to solve, I think there's a component of it where the pharmacies are gonna ask, silly stuff. It could be silly stuff like in their 

policies, how often do I have to take out the trash? And it's, there's 

a little chat bot or something and it says, 

okay, according to policy and procedure seven page two, section six, you have to 

take out the trash two times a day once at night, and then once, you know, at midday or, 

whatever it is, 

and this is your cleaning procedure and so forth. I think in the future as things become more digital and logs start being electronic and 

temperature logs and security logs, and even computer 

and inventory models and inventory reports are all online, what's gonna be fascinating is, a one click audit, Tell me about all my inventory

Losses and 

inventory discrepancies up and down over the last quarter. 

Tell me [00:36:00] how many times, I sent a prescription order and approved a prescription order without a 

doctor's signature, God forbid, 

And read that. When's the last time I had a temperature excursion?

When's the last time I bought 

from a vendor that hadn't gone through blah, blah, blah? How many times did I buy from a vendor that wasn't properly 

licensed in my state or in their own? 

And so, yeah, you definitely have, that's where things are going in two to five years, it's 

gonna take a lot of time Companies have to adopt this new digital way of life, and our future customers have to adopt these technology advancements, and then, 

The technology 

itself has to be built. And 

so that. An LLM or whatever can actually read the data. And so I think, it's kind of crazy because the discrepancy is that, and this might be the case with, something like blockchain or something, but the 

technology has gone so far, so fast

that [00:37:00] the implementation is taking a long time.

 And so, you know, There's so much holding back the implementation, like ev electric vehicles, the technology is, Tesla has proven, 

you 80% of Americans can drive an electric vehicle, 

but the implementation is lacking. Why is the implementation lacking?

Some people love to feel. The gas engine, they want to feel the turbocharge mustang the four Mustang. They just introduced the 

like a, a, 3D kind of feel and 

hearing of a gas engine, but you're an electric car. It's all fake, so 

That's one example. And, the self-charging the supercharging is always a struggle and they're trying to figure that out.

So the technology is there. The AI technology is so far where you could do something like I just 

imagined, 

But the implementation has to catch up. Some pharmacies are still on pen and paper, or 

they're scanning their receding list or their 

purchase list or something like that.

But [00:38:00] it's not smart. They're not mining that data. They're not able to 

leverage that 

data. And so it's the technology providers that have to figure out, okay, how am I gonna implement? Large language models into my solution, 

So much. And then how am I gonna sell it to the customers where they're comfortable with it?

One of the things that we come across every single day is these companies that are like, okay, AI is great. I don't trust it. Why would I trust 

a system telling 

me what I need to do? If I'm wrong, I'm gonna get fired or I'm gonna get in trouble. So what we did is we implemented two things. We actually now explained, look, you need a Kansas Board of Pharmacy, pharmacy license.

You need it because you said you take possession of the drug. You said you dispensed the patients in Kansas. You said you're located in Michigan. You said you handle prescription drugs. So we 

have all the reasoning 

there and then we also why, how we got to that [00:39:00] conclusion. And then we also have the regulatory framework citation based on Kansas.

Co title 

six, section four, paragraph three. This is the exact test. A non-resident pharmacy shall obtain a license before shipping to a patient 

in, in Kansas. They need it. They need it, 

[00:39:22] Mike Koelzer, Host: Sammi, what is your most difficult emotion in your position? Is it worried about having too much power and not being responsible with, there's a million things. What is the toughest thing that goes on in your mind that you might say, I'm kinda lacking this skill.

I wish I was more like this, so that this wouldn't be so hard.

[00:39:50] Sumeet Singh: Yeah, that's a good question. I would say as far as skills go, I started, three years ago, pretty much, to the month at least, we realized. Nathan [00:40:00] is my chief of staff. He used to be operations manager then, and he's the one that turned to me and said, we need a system.

If we're gonna get, if we're gonna two x or three x over the next five years, we need some kind of system. 'cause at that point we had tried everything, we had tried Excel sheets and we had tried X, Y, Z. And he came to me and told me that. I'm like, okay, yeah, let's do it. And so that was the point where I'm like, okay, we need to build technology.

We need to, I need to get out of the day to day. Was a big 

component of it. I 

was doing my accounting in the evenings, every month I was doing the red lines. I was helping with the consulting. I was, solving problems. I was figuring out which documents to fill out. It was a good office disaster.

So I've been incredibly fortunate enough to have built a team around me over the last three years that if I'm not good at something, I just hire for it. One thing I'm notoriously bad at is, I'm not really a an operational details guy, 

but Nathan is right. And our vice president of professional services, Mariah [00:41:00] really is, and she's really good with companies and so forth.

So I would say, so much of my position as as now as CEO of the 60 person company is not doing anything I'm not good at. And giving that 

away, 

giving it away to a qualified, trusted resource, which, I'm proud to say that, my executive team of seven and everything that, they really compliment me in 

ways I never even thought possible.

So 

I think one thing I guess the emotion I struggle with that's to answer your skills question. Then 

on the emotional side it's just this, I don't know what, why I have it, but it's this incredible sense of urgency. 

I just feel like the world is changing so fast and we gotta stay on top of it.

So more than anything, we're enjoying a lot of success. We've 

grown, almost 300% last 

year in, in, technology revenue and we're bringing on customers every week if not 

every day. And so that's all good. But yeah, I think I struggle with [00:42:00] the sense of urgency the most.

It's, am I moving fast enough? Am I doing enough? Because I see the vision, I see exactly 

where we're gonna be in two years.

And yeah, so it's struggling just with patience and the sense of urgency and wanting everything right away.

My, our product management, vice president, product management, Adam, you all, he loves the phase approach, okay? You want we're reports in your system, okay? This is, we're gonna do phase one, it's gonna be a soft launch. We're gonna use it, we're gonna 

send it to our internal team, we're gonna get 

feedback. Then we're gonna do phase two, a light launch, and we're gonna have four,

Template reports. Phase three is custom reports. 

And I'm thinking like, man, can we just skip to phase three tomorrow and no, it's 

gonna take six months to get it. so I think that's the biggest thing I, I struggle with.

but you know, I'm, I'm, grateful for the journey. 

[00:42:50] Mike Koelzer, Host: My brother's a consultant. He owns a brand consultancy out in uh, he's out where you are. He's

20 miles east of San Francisco in Moraga. [00:43:00] There's a city out there called

Moraga. That's where he is.

Anyway, we always joke because in consulting, you have to have all these phases, so his wife will ask him to take out the trash, and his first phase will be, locating where exactly the trash is and which way to walk up to, you know, all this kind

[00:43:15] Sumeet Singh: Right, right, right. 

[00:43:16] Mike Koelzer, Host: You know what's funny to me, you'll read some stuff about pressure and anxiety. It'll say, don't read the news because the news just hypes you up. And I'm like, no, it doesn't. I like to read about the murders and the car crashes and the illegal stuff going on with people, but one place that I've gotta be careful reading is about technology, because I kind of feel that too.

I'll be sitting there some morning and all of a sudden you read about new stuff coming out. And I'm like, I gotta get on this today. Someone else is gonna be doing it. I do have to be careful. I love it, but I gotta be careful about reading about technology because you can feel like you're behind, even though you're one [00:44:00] out of a thousand, you know, you're on the cutting edge.

More than 999 people out there.

[00:44:05] Sumeet Singh: I guess being on the cutting edge of technology and stuff like that, it, a lot of it is just, yeah, maybe reading about technology is probably just a lot of noise. I think. I think what it comes down to is what are the problems that you're actually trying to solve?

And making sure you solve for them. Mike you mentioned earlier this call, like you've done certain things to free yourself up to be able to, pursue other things and to have a work-life balance and to do 

all these things. You went digital first and you did all these things.

There are pharmacies out there that, love the tried and true method. They're on paper. They 

buy, 

by calling on the phone, I need, six 

of this and 12 of that and blah, blah, blah. And so I think definitely there, there's that component which is, what are you looking to accomplish and how are you gonna accomplish it? 

Because the fact of the matter is like there's technology out there in the pharmacy space that is absolutely [00:45:00] incredible. It's like, why isn't 

everybody using this? 

 There's,

purchasing, software like like TRX eight or SV rx or

Something like Sure Cost or something like that, that automates and sure 

cost even creates compliance with the big three wholesalers and, this is how much you're saving on a monthly basis Then there's, the buying part of it and the operations part of it. I think you had Shift RX on last month 

and, they're trying to figure out how to automate and make it more delightful for scheduling pharmacy tech 

shifts and, and stuff like that.

 There's always more and more technology to come in. I think it takes a disciplined buyer. You have to stay disciplined, you have to stay 

[00:45:37] Sumeet Singh: on top of it. There's 

been some big flame outs. I think tabular off the bottom or something, but, there's a company called Prescribed Wellness and 

It, was all the hype that, oh, we can add $30,000 of profitability per year and this is how we're gonna do it.

But when you 

get down to implementing the product, it's not really there. The technology was getting all this hype and it was 

growing so quickly that they weren't able to keep up with [00:46:00] actually delivering value to the customer. 

[00:46:02] Mike Koelzer, Host: Yeah, you've gotta look at your problem, not so much at the tools. You can look at a lot of cool tools, but if you just have to build a box, you probably have everything you need for that problem,

[00:46:13] Sumeet Singh: Yeah. If you're buying, $50,000 worth of products per month, does it make 

sense to spend $25,000 a year 

on, optimizing another two or three percentage points and 

bringing in the net profit or whatever. 

Maybe not, but yeah, 

If you're buying, a million dollars a month as a 

pharmacy and you're rocking and rolling, six states and blah, blah, blah, 

then yeah, definitely makes sense to double click that.

'cause you're leaving money on the table. 

So creating discipline and, vetting out the ROI, making sure the, you look at customer references and stuff like 

 that. Is kind of the way to go.

[00:46:49] Mike Koelzer, Host: You and I we had talked before about blockchain stuff and you gotta be careful as companies you could do a ton of stuff with programming and all that kind of stuff, but you've gotta focus [00:47:00] on what problem are you solving. Probably get there with the most efficient, effective way. And so that, I think that's good advice. Not to get too far ahead of ourself. Just 'cause

the technology's there doesn't mean we have to use it. We can use it when we need it.

[00:47:15] Sumeet Singh: right, when it makes the most sense and there's a clear ROI and you can't tolerate risk or you can tolerate the time or whatever it might be. 

[00:47:24] Mike Koelzer, Host: Sumeet, thanks for joining us. Pleasure to meet you. Anytime somebody is willing to go between me and the government or me and A BBM or something like that, I'll take it.

You can stand right in the firing line. So thank you.

[00:47:43] Sumeet Singh: Yeah, absolutely. No, thank you, Mike. Yeah, it's been a, an absolute pleasure to be on the podcast and, I thought about things I've never really thought about as it relates to, why regulations exist 

and all that fun stuff. I appreciate the conversation. Mike.

Thank you. It was 

a pleasure being on,

[00:47:56] Mike Koelzer, Host: thanks again, sume. We'll be in touch. 

[00:48:00] You've been listening to the Business of Pharmacy podcast with me, your host, Mike Kelser. Please subscribe for all future episodes.