Interview by Audrey Russo and Jonathan Kersting
What happens when AI moves beyond buzzwords and starts helping pharmacists get critical medications to patients faster, safer and more accurately?
On this episode of TechVibe Radio, we welcome LiAna Tschen, VP of Platform Operations at SoftWriters, and Danielle Greer, VP of Product, to explore how the Pittsburgh-based company is bringing intelligent automation into long-term care pharmacy. SoftWriters powers pharmacy operations for facilities where accuracy, speed and scale are mission-critical, with some customers processing thousands of prescriptions every day.
LiAna and Danielle detail how SoftWriters built its AI implementation around one of the most labor-intensive and error-prone parts of pharmacy workflow: order entry. They explain why the company created an internal AI innovation center, how customer feedback shaped the project, and why keeping humans in the loop remains essential when patient safety is on the line. They also share how SoftWriters balanced its core platform work with AI experimentation, spending more than a year testing, validating and building trust before going live.
Listen to learn how SoftWriters is using AI to tackle staffing shortages, margin pressure and growing demand in long-term care pharmacy, while keeping the patient at the center of every decision.
The Pittsburgh Technology Council produces TechVibe to explore Pittsburgh's technology and innovation ecosystem.
Transcript:
Softwriters
===
[00:00:00] Tech Vibe. Here we are telling the stories of what's happening here in Pittsburgh. And Audrey, so much fun when we get to talk about companies that are working on some technologies that literally, they save people's lives. They keep people healthy.
And when you can have AI do that, that's pretty, it's a pretty heavy-duty lift. And today we are gonna be hanging out, of course, with Softwriters. And I have Shen here today. And of course, we have Danielle Greer hanging out with us. Audrey, how much fun is this? It's way fun. Yes. I think we should jump right in.
In the Huntington Bank Studios. Just to set the stage just a little bit, I mean, for, for 20 years Softwriters has been around. And the idea that they're do- doing all this work around pharmacy within long-term care facilities, I mean, talk about a cool niche, a growing...
And I c- I only can call it a niche 'cause there's a lot of people- ... that need these services, and literally life-saving. Yeah, no wonder. And being able to put a little AI into this to help the pharmacist out and to make sure patients get the right meds at the right times and all that kinda good stuff.
We think it's cool. We [00:01:00] love it, man. We do too. We, exactly. We love it. Well, who, let's see who we have in the room. And introduce yourself. Sure. My name is Liana Chen. I am VP of Platform Operations at Softwriters, which means that I am managing the bulk of our product management team. And I'm also a pharmacist by trade, and I was a customer of Softwriters for quite some time.
I love that. Wow, that's great. You like it so much you joined the team. Exactly. That's right. That's a pretty linear path. That's great. Yeah, absolutely. I'm Danielle Greer. I'm VP of Product, and I get to lead what's called our Alpha Lab at Softwriters, which is the actual AI and innovation center doing all the research development proof of concepting, experimentation.
So we get to work on some pretty cool stuff and deliver it to an important population. And how long you been doing that for? Oh, I've been with them for four years now. Okay. So it's been quite the journey because, like Liana, I was also a customer for- No way. Oh my goodness. Oh, wow. All right. I spent almost 20 years as a customer and worked in the long-term care pharmacy, but I've always been on the [00:02:00] technology side.
So I started in this industry with robotics, the remote automated dispensing medication machines and- Yeah. Yeah. Absolutely ... integration. So I, I got to come up in the days before internet and everything. When we were literally still walking and driving floppy disks out to load orders- Mm ... into the machine and stuff.
So it's been quite the transition, so this is great. It'll be great to get up to speed to see what you're all up to and what's changed. Oh. Yeah. Yeah, we're excited. So what, let's, let's just sort of talk about this. When you talk about what you do, sort of peel it back- ... and describe the product. Yeah. So, you know, what we have is a pharmacy operating system, right? So, anything the pharmacy needs to do to kinda do the work of a pharmacy, intake orders, process prescriptions, adjudicate claims, print prescription labels, get those out to the customers. In long-term care there are also needs for, like, automation, like, kind of what Danielle had worked in before.
So the software needs to be able to interface with automation [00:03:00] both in the pharmacies as well as in their customers' buildings, and then also interface with any of the eMart applications that our customers' customers are using, right? So we're intaking and sending electronic information back and forth between all of those different integrated partners.
I know, I'm- It's a very complicated- I'm exhausted. I know. It's a b- Yeah. I think it's one of the most complicated parts of- It is Right ... of the pharmacy industry. Long-term care pharmacy is very, very complicated. And if one thing goes wrong- ... I mean- That's right ... the results- Right ... aren't gonna be always that great, so I mean, that is just, you gotta be precise with every step.
And the patient is at the end of it all all the time. Right. It has to be the center of everything that we do, everything our customers are doing. It is. We have an entire suite of products. So we've got the core system. The pharmacy management system- Okay ... FrameworkLTC. It's an entire ecosystem to take care of.
They say it takes an army, right? To service this industry. I believe it. And it does- Right ... 'cause we do everything from the consultant pharmacist software- That's right ... to the we've got Framework Vision, which is an externally facing software that the nursing home and nurses can engage with. We've got Flow, which is a device that they can wander through their pharmacy and work a little bit [00:04:00] remotely.
Document management- What is that one? ... so like an enterprise content management. Talk about that. Oh. Oh, Framework Flow? Yeah. Yeah, you can get a little, you can work it off your cellphone or a, a device that you can purchase. And there's, there's parts of the pharmacy workflow that require you to be m- walking around, so being tethered to a desktop doesn't always work for those.
For example, toting, getting the medications into the totes to be delivered to the facilities. So- ... we were able to take a version of our software and scrunch it down and get it into a mobile device to- Very cool ... to put those critical workflows in there. Yeah, one of the things that the pharmacy system does that doesn't happen maybe in the old days of pharmacy or even in m- some retail systems is sort of barcode scanning, right?
Both for safety and also the ability to kind of manage your workflow. So that device can kind of help barcode scan without being tethered to a desk. Pharmacists must be loving you. Because they're able to get their job done. We like to think that they are. We hope so. Yeah. But they're pharma. I, I, I know.
Like, how big of an opportunity is this? I know we have an aging population. But do you also- How would you, like, for in terms of like- [00:05:00] Yeah, like is it... You know, I know healthcare is like 25% of the gross domestic product- in the US. I know that nursing homes are struggling with staying open- ... and, you know, providing services. So the demand has to be very intense. So- It's very intense. They're, our customers alone, they process millions of prescriptions a month. Millions. And some of our largest, larger customers are doing, you know, thousands and thousands a day to keep up with- Wow
you know, the demand. Right. You- Like, your busy retail pharmacy is gonna be maybe 500 scripts a day. Okay. Yeah. Like, your busiest retail, like if you were to go to a CVS- Right ... or a Walgreens. Maybe you can find one that gets up to like 900 on some days of the week- Okay ... or some days of the month. But for the most part, 500 or 600 is pretty busy.
So in contrast. Yeah. Our largest ones are doing, you know, upwards of 5,000 a day. Oh my goodness. Yeah. Wow. That's so amazing. And what I love the fact is that your technology is keeping that straight. You're keeping it flowing and going. That to me is just an amazing thing. Yeah. And that's, we don't think about that.
Like, we, we take it for granted that you go to pick up a prescription, it's there, it's the [00:06:00] right thing. Oh, yeah. The label's correct. Oh, yeah. Like, all the things you need to take that medication safely- Mm ... and have it on time, right? Just imagine the nurses- Ha- ... and the LPNs that are in these long-term- Oh, yeah
care facilities- Oh, yeah ... that have all these other responsibilities. Yeah. Pharmacy is just one s- It, one- ... part of what they do ... one part of that. Exactly. Man, that is amazing. Technology's become a requirement to do business at scale for the pharmacies. Yeah. 'Cause they have to process so many prescriptions, so many orders safely in such a short amount of time.
And anytime, that's why we've seen such a shift, too, with more of our pharmacies picking up robotics and- Mm ... dispensing machines internally because they've got to process at scale and volume. And that's why AI is so important. I was gonna say, I mean, this seems like if there's a, a use case for AI, but a really tough use case for AI- Yes
this is definitely it. And so I mean, you leading up the, the alpha lab for this thing, that's some responsibility right there. Yeah. It's- Where you're like, "We're gonna tackle this problem." Feel it. It's gonna be- Feel it. I know. Yeah. And so can you talk about that alpha lab, at least at the [00:07:00] high level?
Sure. Yeah. Like what that, that means. Yeah. We, we we meet with our customers regularly, and we take it very seriously. We listen directly to our customers. 100%. And since I've been with the organization, we have spent a lot of time every year going out, meeting with our customers where they are. We walk their pharmacies.
We get very into the workflow and see what the problems are, and we bring that back. And- About, what, a year and a half ago, we formed what was called an AI mastermind, which was about 20, 30 of our customers that participated in a few sessions- Okay ... where, you know, we really just wanted to hear them and listen to what the problems were and how we could solve them.
And the common theme that came out is, you know, we're dealing with staffing shortages, we're dealing with margin compressions- ... we're dealing with uncertainties. We've got a growing population- Yeah ... and more demand. And we're like, "Okay, so, you know, we've got a problem to solve." And- Yeah ... and then lo and behold, what happened a couple years ago, the AI started to take off more, and it became more accessible [00:08:00] And it became more mainstream, and that opened up a, a bigger door for us to be able to look at these manual workflows and look for opportunities to automate some of those.
And we also decided that we needed to still keep our core product- ... which we decided to bring Liana on board- Mm ... to help keep the core running- ... which is doing all of the upgrades and stuff just to, to keep the train moving basically. And we also decided to carve out a small group of us that could focus on 100% innovating in the AI.
'Cause what happens is if you over-index on one and not the other, you tip the scale. Exactly. And we wanted to keep the balance. Yeah, you got ... There's, there's a balancing act- Yeah ... you have to maintain there. But how cool- How cool ... you could actually then work with your customers on this. Yeah. And then I guess the, obviously they probably helped you beta this as you were rolling things out.
Oh, yes. Yeah. So cool. Wow. How awesome is that? Yeah. We have great customers. We do. I'm gonna tell you. No, they're always so willing to give us information- Okay ... partner with us. And when they found out that we were doing this exciting project, and the order [00:09:00] entry portion when we had that AI mastermind was the thing that stuck out.
So we, we did a, a little scan the barcode survey to find out, like, what's your most critical pain point right now? And unanimously every single person said order entry. Yeah. Okay. It's the, it's the most labor intensive piece- The beginning piece ... of the pharmacy. Yeah. It's also- Is that the first step?
It is. It's pretty much- Yeah. Is it? It's pretty much the first step, and it's also the hardest to train. It takes the longest to train those techs. Okay. Oh. And when you, when you're in a tech shortage, it's just, you just, everything grinds to a halt and everything gets delayed behind that. Yep. Oh, my goodness.
Yep. Wow. And that, that's what they really were, like, "Hey, if you can solve this problem, it's, it's a great start." It has to be very foundational. If something wrong goes there, that's only gonna get amplified down the chain, right? Oh, absolutely. So that's- Safety's important. Yeah. It takes a lot. Like transcribing errors, you don't think about that, but the, the error rates in pharmacies are very high.
Some say it's 25. I can't read the doctor's handwriting. Can you, Liana? Nobody can. Nobody can. But hasn't technology helped with the handwriting piece? It has. Yeah. Yeah. No, a lot of the mandates [00:10:00] and stuff that'll come out- Right ... have to be, like, electronically prescribing. Yeah. That's what I thought. But, you know, you still have some bit of transcribing that goes on in the software to get it into- Yeah.
'Cause, you know, they- ... you have nurses that are putting the orders into- Yeah ... an electronic MAR system at the nursing home, so they're transcribing something else. And then it gets to the pharmacy. So by the time it gets out the door, it's probably been looked at and retyped a couple of times.
So having, having that through AI and being able to automate that workflow step really does that as well. It not only takes some of that manual effort off, freeing up- ... time, but it also increases the safety, and the overall error rate goes down- I know ... because you have a very- The error rate has to be so high- Yeah
on manual. Oh, it is. They s- the, around, I think, 25 to 35%- Wow ... hap, of errors- Wow ... happen during the order entry part. Wow. And what I love is- Your AI, your software is avoiding this and making lives easier for these folks to do this and do it accurately. Yeah. What was the timeframe from, , [00:11:00] you, you're saying this is the problem- Yeah
to solve. You're working with the customers, and now you've rolled this thing out. It felt like... It felt long. Yeah. It felt- I mean, that's a massive lift ... it felt long. We were in it. You know, the, the point was we never really focused on the time.
Okay. We focused on the patient and the safety factor. Gotcha. Right? Yeah. We... 'Cause there were so many options out there, right? We could have taken the route a lot of people say, "Hey, we're automating things," but really it's, it's what we call RPA, which is just plain old robotic- Right ... process automation, which is like screen scraping and- Right
bots that click on things. And we really wanted to legitimately come out with intelligent automation. We wanted to continue with some traditional automation- Okay ... but we wanted to use the AI, the machine learning, and the LLMs- Yeah ... to leverage intelligent automation as well. So we knew that we were building with intent, and it was not gonna be a six-month project.
It wasn't gonna be a two-month project. Yeah. So we worked on this for about 12 to 18 months. And- Amazing ... and we were, we'd been in the alpha phase, we called it. Okay. We were able to get to alpha through partnering with a wonderful customer, and then we expanded [00:12:00] it to a couple of more customers to really get that safety and comfort level of the, and the buy-in of the pharmacy to know that we could run the data through.
We could run it through- Yeah ... at scale. We could see the 99% accuracy. We could identify the more complex orders. So, start to finish, where we finally went live and felt comfortable enough, was about 14 months. That's amazing. Yeah. I feel like you could have, like, you should have, like, captured all the steps.
You have, like, a little masterclass - Yeah. ... of, like, how you de- develop and deploy- Yeah. ... Like, a high-end AI project. Like, that's, that's- It is, and this is a area, like, like I said, we couldn't... You can't go wrong in this area. Exactly. Yeah. We, we were very focused on- Yeah ... ensuring that we had the trust of our customers, that we were building this right, and we had the right- Oh, goodness
the right goal in mind. Yeah. I'd be freaking out, like, oh, my God, oh, my God. Right, and as a pharmacy- And we still freak out, yeah. Exactly. Yeah. It's really important, right? That we're not moving anything forward in the process if we're not sure that it's the right thing. So, you know- Got it. Yeah ... we don't automate everything 'cause there's a lot that- Of course, yeah.
Ironically, even [00:13:00] though, I think, you know, it took us a while in long-term care to get the bulk of the orders electronically rather than on paper. There's still quite a bit of paper. Okay. But even those electronic orders, still there's a quite a bit that's not really structured information, so that there is quite a bit of interpretation that we have to train the models to.
And so we really wanna make sure we get that solid, right? Makes sense. And that we're passing through. It, it's not making it through that step if we're not sure it's right. And the medications are changing all the time. And you al- also in other long-term care, you talked about prisons? Or- Some f- long-term care pharmacies service prisons or jails or that t- Okay ... those types of settings. Yeah. Wow. Where you have constant turnover too of patients. Those are the that's its own challenge. Oh, interesting.
Right. That makes sense. Yeah. It's not like you're living there- in your last chapter. Yeah, especially in like a jail setting, right? Right. Come in, and they leave- Right ... quickly. Right. Wow. I'm just mesmerized at the whole scope here that, that goes through to make sure people get their medication.
Like, I just feel like, you know, we, we wait in line at the pharmacy. You pick up your stuff. You're not thinking about that, much less when you're in a long-term facility where they're managing all these different [00:14:00] patients' prescriptions where there could be 12 of- Wow ... on average. And what I think really is interesting, Audrey, is the fact that you're not saying, "We're gonna automate the whole thing." You actually are just automating the parts that you know it can be done best. Well, it's gonna have the most impact for the customer ... the most impact for the customer.
Yeah. And then, and there's always apparently like, like a human in the loop here as well too- ... of course. Yeah. And I'm thinking so the pharmacist now can really concentrate on some of the more important things, not the tedious kind of back and forth stuff Well, what's it like for pharmacy grads these days?
Are we getting a lot of people or a lot of- interesting. I would say we don't have as much as ... There used to be a shortage of pharmacists, not quite as much anymore, although I think we might be running back into that if people aren't going to pharmacy school as much as they were. That kind of ebbs and flows- Okay
and might vary regionally- Okay ... in terms of what that market is like as a new grad. I I think pharmacy technician, which is, you know, the normal- Different, yeah ... workflow is, you know, a technician might enter the orders or they may fill, like they might fill- Mm
the package. And then a pharmacist checks their work. And so that pharmacy tech role, I think, is where the biggest crunch is, right? Ah. I [00:15:00] think after post-COVID there was the most I think we lost a lot of workers, and then wages went up, right? The demand for wages. Right. Right. Yeah. And then at that same time, now we're in an era, era where I think pharmacies have always been worried about reimbursement rates, there are more economic pressures on the, the pharmacies right now. Yeah. And particularly long-term care pharmacies. When you think about, like, all the, anything that's going on with Medicare Part D- Right. ... cMS, all of those are gonna put a lot of economic- Mm ... pressure on the long-term care facilities, and then that, those are their customers, right?
Right. So they're being pressured to give better contracted rates. Yeah. And when we spoke to the customers back when we did the AI mastermind, it was, I think these were some of the exact words in talking about the staffing shortage the, the pay scale not being able to keep up with industry. Can't increase the rates that way because their reimbursement rates- Right, because their margins are thin
aren't matching- So they have to be able to- ... to be able to go up ... they're, they're tapped out. Right. That's, that, that's the best they can do. That's right. So, so basically that's why they're all looking to us, right? They're like, "Solve the problem. Automate as much as you can." Right. You mentioned we picked the, the one thing.
But we just got started. Well, now- [00:16:00] Yeah. ... you're reading my mind here, 'cause I knew. I had a feeling Audrey, that they're just getting- Yeah ... cooking on this thing right now. Yeah. Also- We just got started ... so what's next? Yeah. What, what's in the alpha lab right now that you guys- Well, and part of that exercise that we've taken over the last, you know, year or two is really just looking at that entire value chain, the start to finish of an order.
Okay. Yeah. And how every step that it takes to get through a pharmacy and eventually make it to the patient. And there is a lot of opportunity there. Yeah. There's so much space. There is. Yeah. There's a lot. 'Cause there's how the order comes from, you know, the originator, right? The doctor, whoever it is.
The nurse types it in. It makes it to the pharmacy. It gets put in. So that's that order entry piece that we decided was the most, you know, critical. That's what our customers told us to start with. But there's a whole nother part of that. There's how the pharmacists verify. There's that PV1. That has a lot of opportunities- Okay
for us to create what we call, like, a co-pilot or an assist, an AI assist- ... in that area to help them. This is so great. Man, amazing. But how much fun- So they're in Pittsburgh. Yeah, yeah. Started in Pittsburgh, pharmacy. Ain't leaving [00:17:00] Pittsburgh. Seeking for error- No, we're not leaving Pittsburgh. Seeking for error free, right?
Yeah. Innovating each and every day. New products down the line. Yeah. And you're hiring. Yes, we are. Looking for some engineers. Yes, we are. Great story all the way around. It's a great story. And how much fun we get to tell a little bit of that story. Our front row seat- Have the fun to listen to them
here on Tech Five. Yeah, that's right. Listen to them. Fantastic. Encourage everybody- And we're excited to get the message out, too ... keep, absolutely. Yeah. Yeah. Keep coming back for more of this and- Yeah. Softwriters.com? That's what it is. Yep, yep. All day long. That's great. Come see us. Simple as that. Yeah. Thank you.
Yeah, thank you. Thanks for hanging out with us today. Thank you. What a fun way to spend part of our day. Yeah. That's for sure. Yeah. I learned so much, and once again, you guys are making Pittsburgh proud. Simple as that. So, dude, too much fun. It's great. I cannot wait to see you guys on the next one. Yeah, thank you.
All right, sounds about right. All right, everybody. Thank you so much.