This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

[00:00:00] This episode is brought to you by SHI. As a trusted partner to healthcare IT leaders, SHI understands the unique challenges of managing complex healthcare technology environments.

Their IT asset management team has helped health systems save millions in licensing costs, eliminate security vulnerabilities, and optimize IT investments. Whether you're modernizing your infrastructure, implementing AI solutions, Or strengthening your cybersecurity posture. SHI expertise helps CIOs and their teams focus on what matters most.

Enabling quality patient care through technology. See how other health systems have transformed their IT operations at ThisWeekHealth. com slash SHI.

Sarah Richardson: Welcome to This Week Health. I'm Sarah Richardson, a former CIO and President of this Week Health's 2 2 9 Community development where we are dedicated to [00:01:00] transforming healthcare one connection at a time.

Now onto our interview

(Interview 1) Welcome to another this week, health Solutions Showcase. Today we are excited to feature SHI International, a leader in empowering healthcare organizations through innovative AI and cybersecurity solutions.

Joining us are two distinguished technology leaders from SHI lee Ziliak Field Chief Technology Officer for Advanced Growth Technology at SHI who will share insights on i's AI lab capabilities, partnerships, use cases, and the broad range of services they deliver.

And also Kris Nessa, CTO for SHI healthcare who will validate how SH i's AI expertise is helping healthcare providers navigate and accelerate. Their AI journeys with real world success stories. Together, they're gonna show how SHI's AI and cyber labs provide a risk-free environment to innovate, validate, and deploy AI solutions at scale, all while maintaining a strong focus on measurable outcomes, security, and operational excellence.

Lee and Nessa, welcome to [00:02:00] the show. Thank you for having us, Sarah.

Lee Ziliak: Yeah, great to be here.

Sarah Richardson: Oh my gosh, I love this topic because you have empowered organizations to innovate and validate AI solutions in a risk-free environment, ensuring they achieve measurable ROI before full scale deployment, all with trusted guidance and expertise from SHI.

If either of you wanna jump in first and share more, love to hear it.

Kris Nessa: What's funny, and thank you for asking that, Sarah, I actually had pulled over our CEO Ty Lee at the SHI ribbon cutting the other week.

I had a brief moment. We were walking the hall together, just the two of us. And I complimented her. I said, I want you to know, I think you're doing the right thing at the right time. And she ironically was a little nervous. She's like, oh my gosh. She really thinks so. Like she's still trying to find that balance maybe herself.

cause I said, I've worked at one other place that did the right thing at the right time, and I think you're doing it too. She said, tell me the story of the first one. So very long [00:03:00] story short, when I used to work at Cerner, I think we were doing the right thing at the right time when we started doing remote hosting and allowing customers to ship off their it onto SHI to Cerner, two different business verticals.

So Cerner being electronic health record company all of a sudden spun up remote hosting and. Manage capital of an IT. Like you see EHR companies doing that nowadays or trying to do it. But back in the day when Cerner did it, it was a game changer. And Cerner could do it cheaper, faster, more secure, and better than a lot of hospitals could do it at that time.

And it blew the business up. Well here you got SHI at the right time with AI almost doing the same thing. We have stood up and verticalized remote hosting and or bring your idea to SHI first. It is the best playground to come and bring your data, bring your idea. We have software engineers, we have data [00:04:00] scientists, and you can spin up an idea into a prototype and try it on different storage systems, different Nvidia chips and all that.

Before you ever sign a check to buy any of that infrastructure, can you imagine spending 7, 8, 9 figures on this stuff and bringing it home and then not knowing what to do, or you bought the wrong stuff? Like we're here for that playground for you to find and get the most value before you actually spend that money.

And I think it's just genius. Honestly,

Sarah Richardson: it is genius that safe space is gonna be huge because you don't know exactly what you're paying for yet in some spaces. And so, Lee, this is such a sweet spot for you when you think about where organizations can position the lab to address their concerns, have insights, and really streamline.

Both the operations, but to Nessa's point manage the cost.

Lee Ziliak: Yeah. So let me back up before I address that just a little bit. What that ribbon cutting was specifically on our AI [00:05:00] and cyber lab. And so we've been doing AI services, this experimentation for.

Almost two years now, but we had our official ribbon cutting on our on-premise lab about three weeks ago. And so what that lab does is exactly what we talked about, right? It gives you the customer a safe place to experiment with your AI use cases. It is really. Proving ground that we leverage to make sure that your use case is valid, that you have the data to support the use case, that the models work the way that you will expect them to work, right?

Or we will expect them to work. And to her point, I. You can do this without a nine figure investment. So that is the function behind the lab. And like I said, we've been doing that for a few years and the lab is equipped with about 60 GPUs right now. So, again a variety of GPUs, but also a variety of platforms from different OEMs.

So we've got obviously [00:06:00] Nvidia. Equipment, but we've got Dell equipment, we've got HP equipment, we've got super micro equipment. So whatever your target is. We can help validate that. And if you don't have a target, then we'll help you figure out what that is as well. So, she mentioned a nine figure investment that's not uncommon in terms of, if you look at like a soup to nuts type AI project.

I gotta go from ideation, I gotta spin up the expertise. If I don't have it, I have to go out and hire it. I have to put a project together. I have to buy hardware to run it on, and then I have to deployment and that can take literally months to years. And so, another thing that we do differently than a lot of companies and even a lot of consulting companies, is we'll help you iterate through that in six weeks.

Our typical engagement is a six week POC. We'll bring our AI team, data scientists, ML Ops engineers, application developers project managers, et cetera, to help execute that POC. Within that six week timeframe. And the other reason [00:07:00] that's important is there's Gartner stats floating around out there that literally 80 plus percent of AI use cases fail at some point, either from inception up to production and again.

If you've made a nine figure investment, then that's pretty painful. So that's what we're there to help with. We help our customers do that across a wide variety of use cases and verticals.

I.

Sarah Richardson: one of the case studies that you highlighted from this lab was entity matching using generative AI.

One of your healthcare customers had been using basic fuzzy matching and manual effort to match records to ensure correctness and completeness of data, and that fuzzy matching was allowing to automate about 20% of their records, and then the rest of the 80% was manually matched and you built a solution that brought forward the power of LLMs, created an application to review the matches, and built a pipeline to continuously monitor [00:08:00] the models for performance and refresh that model embedding them on a regular basis.

Can you share a bit about that value that was delivered and what it was like to work with that organization?

Lee Ziliak: So this is a really great example of what I'll call the business side of the world, right?

So when people think healthcare and clinical work, they're thinking, oh, well. We're looking at images and we're helping do diagnostics and all that other stuff, and that is very important and we can certainly help with that, right? The flip side of that is you gotta keep the lights on, right?

And you have to do that as efficiently as possible. And the, so this is a great use case where we took what was a very highly manual process. And leveraging AI provided automation to help that, that workflow speed up. I'm not familiar exactly with the savings amount, but it was pretty significant in terms of just manual effort and whether that translates to savings or not, it translates into other productivity, right?

So, one of the big fears about AI is I'm gonna lose my job. [00:09:00] If I automate something with AI and that is, one way a company can go, but what I normally see is they'll give you more work to do, which you may or may not like, but if you're more efficient, then you're gonna be able to be more productive for the company.

And this is a great use case example of that.

Kris Nessa: we talk a lot even about AI and a lot of the first use cases that are prevalent out there is like the clinical transcription and dictation and things like that everybody's talking about.

Right. And we always speak about allowing our clinicians to operate at top of licensure. But there's actually, to Lee's point, there is the flip side of operational top of license and top of position as well because we're dealing with staffing shortages across the board in any of our friends networks and all of our provider organizations.

There's plenty of our friends in rural organizations that are barely surviving nowadays and they're the only healthcare entity from miles. Right. For some of our friends and family. And so that [00:10:00] operational side, revenue cycle, helping everybody over there like with automation and allowing like the single person who's doing coding or the single person doing claims to all the

sudden, like expedite a lot of manual process and start to automate things with ML and AI and allow them to spread their wings and do other important things that the organization needs is a huge cost savings and a huge benefit as well. And I think we do need to talk about that a little more too, other than just the clinical side of the house.

Sarah Richardson: Well, for sure, because I'm curious when you think about your engineering team specifically and the technical expertise that y'all bring to the table. Tell me more about how that's a differentiator for healthcare clients that are exploring AI solutions and very specifically within that clinical setting.

Kris Nessa: I'll take this one. So, having lived the life a little bit and even running an innovation team myself, like when Covid hit along with education of staff, like you saw a lot of innovation teams disappear from health systems even [00:11:00] the big ones. So we're kind of dealing nowadays with everybody still looking at their budgets and trying to operationalize it.

They have a flat budget. They maybe had to cut budget, had to cut staff in their IT departments, and a lot of CIOs and our friends are like I don't have anything to be innovative anymore. I don't have a budget, I don't have humans. And so what I really like, what we've done is we've stood up. What I used to have there are software engineers who can code an app.

We can code a fire app, we can do anything. And we have AI and ML experts as well. We have data scientists. We have the team that's sitting here to help you do that, even if you cut them a while ago. So why not bring an idea here and you could do it cheaper. You can actually prove out if your idea is something you really wanna do or if you should scrap it yourself.

another cool like use case scenario was actually SHI's [00:12:00] first customer that came in to try to use our AI labs. Can't name the customer, but they came in fully ready to spend that nine figure, paycheck on all the latest and greatest technology, hardware, and software. And when we got to the point of, okay, let's look at your data.

Let's go over it, see if it's ready, and let's get it ready to try it. On the different OEM things that we have in the lab, we actually were able to tell them. Your data is not ready to go do any ML or AI on. We actually saved them and stopped them from spending nine figures with us and told them their data wasn't ready.

Go home, do A, B, and C and let's come back and revisit this a year from now because we truly value like wanting to do the right thing for you and truly ensuring you're getting the value out of what it is you're about to spend, and then the journey you're about to go on.

Lee Ziliak: And I'll, if I can just add a [00:13:00] little bit of color to all this, right?

So even if you have data scientists and ML lops engineers and all that stuff, right? You focus them on your core business, right? Which is probably the diagnostics piece of things. So there's this massive backlog of use cases that we can help knock down as well. But most people don't have those people on staff.

And the reason is, one, they're very expensive, and two, they're very hard to find. So if you had them and got rid of them, or even if you do have them or if you've never had them there's just such a huge demand for this skillset that we're very happy to be able to bring this to the table.

And the example that Kris just gave you was a company that's got a whole bunch of data scientists and ML ops engineers, and. AI and ML people, and they had this use case sitting over here in the backlog. And so even though they had this massive skillset on staff and they've been doing this for 20 plus years back before, AI and [00:14:00] ML was cool, they still didn't curate their data properly.

They didn't have the corpus of data to support their use case, even though they said they did. And again. I'm confident we saved somebody their job, but it was literally a nine figure investment that they were ready to make. And so, but they'll

Kris Nessa: be back.

Lee Ziliak: Oh, they'll be back. In fact, we're helping them curate that data.

The purpose of the lab. If you're, I can't get this across enough. It is a safe place to do your experiments. Yeah. It's not free, but it's a lot cheaper than nine figures. So, and that's the purpose of it, to help our customers iterate through that stuff quickly.

Sarah Richardson: So once they go back and they do data cleanup and they have the normalization, and they have the ability to actually have some sets that are going to bring them forward into the next phase of their AI implementation, how do you help them also with data privacy and security in the development of their solutions?

Lee Ziliak: Well, so, our AI lab that we re refer to as actually referred to as our AI and cyber lab because it is [00:15:00] built in conjunction with our field CISOs. So the lab, the AI portion of the lab is built on our existing cyber lab. And so as we go through and architect the solution. Not only is your data safe within our environment, we will help you scale that into production as well.

Not just the AI piece, but the entire architecture stack. And so security I've been saying this for a hundred years 'cause I'm not old. Security is everybody's business and it's very important especially in the healthcare space. With all the different regulations and potential problems that you can get into that we look at these workloads holistically and help our customers secure those workloads.

And that's, there's a million different ways to do it, but, at a very simple level it's end to end encryption. It is monitoring the environment, it is taking that telemetry and watching where the data is being used, how it's being used, et cetera. So. The solution that we will provide include in includes and encompasses all that.

Sarah Richardson: Tell me more too [00:16:00] about the training and support that goes along with that to ensure that the providers can effectively utilize their AI solutions, think about what's next, create the next version of some of the tooling that they need. How much do they learn from you in those moments versus coming back to the lab to keep maturing their capabilities as well?

Lee Ziliak: Yeah, it's not really moments. What I will say is, it's an interactive process from the inception, right? So we, part of that six week POC is sitting down and understanding that workload and that involves talking with our customers. Then we'll go out and choose the right model and work with our customers as we start to do.

The testing and validation of that model. And then we will work with them in the final sprint to go over that model and how it executes and how it runs. And if they're fine with taking the ball from there we can hand that MVP over. But the next phase, which is an additional, sprint or sprints would be let's roll this into production.

Let's help you understand all the infrastructure that's required for that. As well as [00:17:00] the application that's developed. So there's runbooks and engagement to hand over the application and the infrastructure.

Sarah Richardson: And Nessa, you mentioned some of the clinical efficiencies that are available out there and really thinking about that.

I'd love to hear more though on about your thoughts, specifically on improving patient care and the operational efficiencies for healthcare providers. Like when you think about the ability to truly be innovative in utilizing these solutions, what does that mean for patient care?

Kris Nessa: Ooh, great question. I've always that person that thinks the word innovative is like a mindset and like a mantra in a way of living truly for like an organization and or a team.

And so I really think that there's ways to be innovative to your own point, operationally and clinically and whatever. A lot of key things that are popping up, let's say. Ambient listening, clinical dictation aside. Revenue cycle is a huge area and a huge topic of a lot of our friends. And how do we better operationalize, [00:18:00] expedite claims?

Prior authorization, adjudication returns, DNFBs, just everything over there on that revenue side because they're short staffed over there too. There are a lot of discussions right now on imaging, whether it be academic medical centers, whether it be true huge imaging centers, or even like at a pediatric level.

Imaging is a big thing. Our AI lab is built on some of the preeminent OE EM second help. With like AI ML for imaging. And so that's a use case that's coming up a lot right now is we have staff shortages when it comes to radiology technicians and people reading it. But if you pause and you take your own patient perspective, when we go in for anything, for radiology, oncology or whatever, none of us wanna wait more than even four hours to hear a result, let alone we're usually waiting a week.

Right. So our customers, the healthcare [00:19:00] organizations know that and they hear that from us and they're trying to figure out how can we expedite technology again to work faster, harder, maybe spot some things, elevate something up to the human faster to get a read back. So we're going through some of those use cases right now.

Also the whole digital front door. Patient engagement is top of mind. We have a few customers that we're talking to at this moment with their front door patient engagement chatbot or interactions. Some of them had maybe developed their own gone to market. They've kind of been out there on their website, used a little, not too much, but now they're totally rethinking it.

They're like, Hey. We've got this. Can we bring it into your lab now and totally enhance it more? Can I put this data that I have behind it and let's even enhance it more and see if we get more traction? So patient engagement, digital front doors, imaging's a big one, and clinical revenue cycle, and just [00:20:00] the revenue cycle processor.

Top of mind right now.

Sarah Richardson: So if I bring all of this to you, I'm listening to this and I'm like, dang, I could not only have them help me clean up my data, create the right use cases, get my team trained up, et cetera. Can you just walk us through what that process and timeline looks like? Lee, you mentioned the six week POC, but I'm just gonna call you after this and say, here's all of my ideas and all my stuff and all my people.

What? What is that engagement actually going to look like for a person who wants to make traction in this area and knows they need you in the lab.

Lee Ziliak: So it kind of depends on where you are, right? If you have a hundred use cases we can sit down and Yeah we'll help you narrow it down, right? There's priorities.

So if you have a hundred use cases, we can help you work through that and figure out what the priority is and then start knocking him out one at a time. We can do more than one at a time, but generally that's the approach. So that we'll sit down and say, all right, here's your a hundred use cases.

Here's the funnel. Let's start here what's gonna gimme the best ROI or what's gonna gimme the [00:21:00] most impact to my productivity? What's gonna gimme the most patient satisfaction. There's not a specific KPI. Each use case is gonna have its own set of KPIs and we have to figure out what's important to your business.

So that's typically, a week or two to, if you're, if you've got a whole bunch of stuff, then we can help figure that out. And then we would go straight into the the the POC process, which is a six week engagement. It comes in different sprints. At the end of that, you have an MVP.

The engagement after that is scaling up into production, and that's basic, typically a six to 12 week engagement. And so, you can really have an operational system in probably worst case, 18 weeks, maybe a little bit more depending on if you need data cleaned up upfront.

If you need. Help with that, use case prioritization. But 18 weeks is a good rule of thumb to go from idea to production. So that's kind of what that looks like. But you know, if you're, if you've already got your use cases, then you [00:22:00] know, we don't have to go through that ideation piece, right.

So. That's kind of what the engagement looks like and how do you do that? Work with the context that you're familiar with, right? So reach out to Kris or reach out to your account team and they'll engage me or my team to come in and kind of help figure out where you are and where you need to help.

That's one of my slides. If if I were doing slideware for you, it's like, where are you on this journey, right? Are you, do you not know how to spell AI or are you very advanced? Right? And so. We've got all sorts of things to help meet you kind of where you are.

Sarah Richardson: It is interesting you mentioned though, the very advanced aspect of all of it because I have to believe that AI is moving so fast.

You might believe you're advanced in one aspect of it, the keeping up with it is one of the harder elements. So post lab, you're in your facility. You've got the partnership, et cetera. What is life with you post lab look like to keep that team moving forward? To do the resource allocation, which may mean that we actually didn't need as many [00:23:00] scientists because we can now allocate them to other parts.

How are you helping organizations balance the AI structure, the governance, the ongoing maturation of the capabilities within an organization?

Lee Ziliak: So those are typically ongoing engagements with us. And when I say engagement, that doesn't always mean paid by the way. So, there are consultative only engagements like, AI workshops and briefings to kind of keep you in your team up to speed with alright, here's the GB 200 and why, how is that different from the H 200?

And do I need that? Right? So there's different briefings that we can potentially help with as well. There are. AI governance workshops that we offer for companies that need to, set up a governance team to help the pipeline of AI workflow that comes through there

so there's just, security engagements we can engage again, kind of meeting you where you are for whatever the requirements would be. And a lot of that is Kris and her team. So, they're the front door to that. And they'll engage the [00:24:00] different teams in the background to help bring to the customer what's needed.

Kris Nessa: Yeah, blog posts, webinars, SHI summits. We've got a lot. But again, it can also be just even as a CIO if you missed it or your IT director did just shoot me an email, gimme a call or call up your SHI account exec and we are always here to help you.

Sarah Richardson: What future developments can we expect from your lab?

And what are each of you most excited about now that you've had time to spend with clients in the environment and really seeing, I think, the art of the possible, which is one of the hallmarks of what y'all bring forward.

Lee Ziliak: So from a lab perspective, it is literally evergreen. If I told you what's in there today, it will change tomorrow.

And that's the cool thing. And we continue to. Add to our portfolio of OEMs and resources that we have in the lab. We're looking at, NVIDIA's latest footprint, the NVLs bringing those into the lab. We're scaling up [00:25:00] our data infrastructure. We're looking at doing instead of just InfiniBand we're talking about doing rocky.

So we can show alternatives and connectivity to our customers. So, the landscape there is just nonstop and it's a full-time job just about keeping up with it, much less getting it into the lab. What I think is the coolest thing about not just engagement with customers, but with my job that comes from engagement with customers is.

I've never said I'm the smartest guy in the room, but I got a pretty good idea of stuff that you can do with AI. What's really cool is when we sit down with our customers in one of these AI workshops, and I have 26 years in telecommunications and so I, I think I know that business pretty well and I can ideate on things you can do in healthcare or construction.

But when I sit down with a customer and they say, Hey, can you do this? And I'm like, well, yeah, I never would've thought of that, but I think we probably can. And so the ideas that customers come up with that we can use AI to solve is what [00:26:00] really keeps me coming back. I'm a technology geek, but I love solving problems and I love hearing customers get excited about AI and how they can use it to solve their problems.

Sarah Richardson: That's a great perspective, Lee thank you. Nessa?

Kris Nessa: It'd probably be the same thing 'cause now I'm working with other people too that are excited to just. Solve a problem in the art of the possible, right? Like, like I said, I've got the perfect playground now sitting behind me, and I've got software engineers, and that's all I ever needed.

I've got plenty of ideas rolling around and either I'm willing to help a customer and give them more feedback and more ideas on top of it to take their business idea. But it's just like having the humans and the capital and the space and the freedom to do it and just play with it. And it's like. How do we move the industry forward?

Like how do we help patient care and drive patient values and keep organizations in the black or get them out of the red and keep them in the black like. There's just so much to do and it's still that [00:27:00] balance of people, process, and technology. So how do we, again, maximize the value of the technology and make it like just elevating your humans and your processes that we are still continuing to deliver the best care because.

My kids got a surgery next week, and when I set foot in that hospital, man, my heart's gonna be racing for my little kid. And I want everything to function and be ready and to go. And so like this is just at the forefront that we still have a long ways to go, but there's a lot of great people here to do the right thing and our heart's in the right place still.

Sarah Richardson: There definitely is. The team that you've assembled to work on these has the background that's needed to your point, to figure out how to make things happen and make them happen at scale, which sometimes is where we get lost inside of our own organizations, as most of us having worked in hospitals and healthcare settings before the bureaucracy gets in the way of just being able to have.

Fun with solutioning, things that are very much part of everyone's lives. Healthcare and our personal and private care is a big [00:28:00] deal and how it affects our families now. Being able to be a part of that influence is really special. Thank you both for taking the time to share about your AI Center of Excellence and Next Generation Lab that's empowering people to explore the latest AI solutions, validate use cases, and confidently deploy new technologies

at scale. With help from their experts you can easily accelerate innovation and integrate leading AI solutions across your entire organization. I can't wait to see what else comes out of this lab and the real time application of the wins you are delivering across our industry.

Kris Nessa: to be here.

Great conversation. Thank you.

Lee Ziliak: Yes, thank you.

Sarah Richardson: Take care. Thanks for listening to our Solution Showcase with SHI. That's all for now.

Thanks for listening to this episode. If you found value in this, please share it with a peer. It's a great chance to discuss and in some cases, start a mentoring relationship. One way you can support the show is to subscribe and leave us a rating.

If you could do that, we would appreciate it. Thanks for listening. That's all for [00:29:00] now.