 Applied from Las Vegas. It's theCUBE, covering AWS Executive Summit. Brought to you by Accenture. Welcome back everyone to theCUBE's live coverage of the Accenture Executive Summit here at AWS Reinvent. I'm your host, Rebecca Knight. We have three guests for this segment. We have Dan Sheeran, the Director of Global Accounts at AWS. Thank you so much for coming on the show. We have Ken Schwartz, Vice President Enterprise Analytics at Health First. Welcome Ken and Shale Jane, Lead Data Business Group in North America. Accenture, thank you so much. I'm glad to have you all here. Good to be here. Yes. So we're talking today about driving digital transformation via data and analytics. I'm going to start with you, Ken. Tell us our viewers a little bit about Health First as a business. Sure, Health First is the largest not-for-profit health plan in New York City. It's a 26 year old company. It's owned by 15 sponsor hospitals. So the business model is a little different than most health plans. The sponsor hospitals who own us, we actually share risk with the sponsor hospitals. So if our members obtain their medical services at sponsor hospitals, we have the same goal of keeping them out of the hospitals, essentially. And we, the revenue stays within the healthcare delivery system. So it's a little bit different business model. We've been very successful. We're a very local plan. So we have a big footprint in the communities, the very diverse communities of New York City. We're kind of part of the fabric of New York City. And that's really very much part of our brand. So your patient population is mostly, I mean, who are, who comprises? 1.4 million members, 1.4 million people mostly in New York City. So we like to say, if you ride the subway in New York City, it's very likely that one in eight people are health first members. One in three if you're in the Bronx. Mostly underserved populations in a lot of cases. And people that really, like I said, sort of the real fabric of communities in the city. So what were the reasons that Health First embarked on this data transformation? Really just, again, a 26 year old company kind of outgrowing its infrastructure and really wanting to make sure that we can keep up with growth because we've been lucky to grow steadily over our entire history. And at a certain point in time, the legacy systems and legacy data systems don't support the new ways to do things, prescriptive predictive analytics, some of the great new capabilities that you can do in the cloud. So it became really important to get off the legacy hardware, get off the legacy approaches. And a big people change management to make that happen. I mean, that's kind of what we've been living for about the last three years. So what were some of the goals? The goals are just to be able to do things at scale for, in the legacy systems, I think we really didn't support analytics across our entire membership and our entire 30 million claims a year, 1.4 million members, 37,000 providers. So just being able, initially, just being able to query and do sort of business intelligence at scale across that much data. The old infrastructure just didn't support it. From there, we've gone into launching our data science platform and things like that. So like I said, just being able to keep up with the times and provide more information, get to know everything we can possibly know about our members so that we can reach out to them in better, more effective ways. So, Shale, I want to bring you in here a little bit. How did Accenture partner with Health First in helping it achieve this goal? Yeah, so we work with companies like Health First all the time and you almost have to embark on a journey that starts with a concept, almost the imagination, if you will, and then you take it into a test mode, a pilot mode, and a scale-up mode. And we were fortunate enough actually to be involved in the journey that Health First has had all throughout those stages, if you will. And it's been a very rewarding experience because Health First is one of those companies that actually took a very early lead on moving to the cloud, moving to the new data architectures, and actually trying new technologies such as we recently finished a knowledge graph project with them as well, which is relatively new in the space. So it's been a rewarding experience for us as well. So what are some of the challenges that you faced along this journey, organization of lead technically, and how did you overcome them? I think early on it's whole new roles and new technical paths that just didn't exist at the company. So Accenture, being partner, good support from AWS really helped us. So we didn't have machine learning engineers and data engineers and cloud practitioners. So you don't grow that overnight. So having professionals come on graph as well. We oftentimes you start off with the use case and you have somebody just download things and get going, right? And that's great, but that doesn't really land it. So getting professionals who have done things in the new environments on board to help us out was really key. In the challenges side, I really think the people change management can be really hard. Again, if you're sort of a brand new company or startup and you have to do your business on the cloud and it's dependent on that from day one, it's a lot different then. We have a lot of people, our company's been successful for 26 years. We have to look to the future to make these changes, but we've been doing pretty well sort of on our legacy platforms and things like that. So it's not always easy to just get people to change streams and say like, hey, you really should be doing this differently. So I think the people change management, realizing you have to kind of sometimes lead with use cases, lead with pilots, lead people by the hands to get from point A to point B. Was kind of surprising, but we've learned that that's true. So Dan, you had a nice shout out from Ken here about giving you some props about AWS and what you bring to the value you bring to the table. What do you make of what he said about the people change and how that is, in a lot of ways, the hardest? I couldn't agree more. In fact, that was the first point that Andy Jesse led off with this morning in his keynote, that in any of these projects, if you don't start with leadership that is both committed to the change and coordinated among themselves, then you've got no chance of success. Now, that's a necessary condition. It's not sufficient. You do need to drive that change through the organization. And the scenario that Ken described is very common and what we see in that you start with enthusiasts typically that we often call builders who are going to be at a department who are playing around with tools because one of the advantages, of course, of AWS is it's all self-serve. You can get started very easily, create your own account. But it is tricky to make sure that before that gets too far along that an enterprise-wide architecture and strategy is agreed upon or else you can get sort of half-pregnant with an approach that really is not going to serve the long-term objectives. And that's the reason why working with Accenture, getting the reference architecture for a data lake really agreed on early on in this project was essential. And that's what allowed once that foundation was in place all these other benefits to accrue pretty quickly. So on a project like this, how closely are you all working together and to get the job done? And what is the collaboration? What is the process and what does it look like? Well, I'm sure that each of us is going to have an answer to that. But our perspective on that at AWS is to always be customer-led. We have some customers who themselves want to use a journey like this to become a builder organization. And one of their strategic objectives is that their developers are the ones who are really at the controls long-term building out a lot of new features. We have other customers who really want to be principally buyers. They'll have some enthusiasts here and there in their organization, but they really want to principally define the objectives, participate in the architecture, but then really lean on somebody like an Accenture to implement it and to also stand behind it afterwards. So in this case, Accenture played a central role, but we really think that the very first meeting needs to be sit down and listen to what the customer wants. Yeah, I'd say we're builders, but with guidance that against some, we want people who have hit their heads on things and kind of learn from that. And that's the force multiplier instead of having. And we definitely jumped into use cases that we wanted to just build. Like I said, and a year later, we're a little bit spinning our wheels. It's not really hurting anything because it's not necessarily anything anybody else is for anyway. It's standing up a graph database. It's just something we wanted to do, right? So having these guys come in as force multipliers has been really useful. So we reach out to AWS, have really good support from AWS. When we need it, AWS also has great online training, the loft in Lower Manhattan, or in SoHo we go to things as well. So we can help ourselves. And then Accenture has just really been embedded with us too. We have seven or eight data engineers that have really walked pretty much every mile with us so far on this journey. So, yeah. The only thing I would add to it is that we have a very strong relationship with AWS. And as such, we become privy to a lot of the things that are coming down the pike, if you will. So that can add value. At the same time, we have very good access to some of the top technologists within AWS as well. So we could bring that to bear. So that all kind of works really well together having a partnership with AWS. And then with our, we have different parts of the organization that can also bring not just the technology skills, but also domain skills as well. So we can add to some of the thinking behind the use cases as well. So that's another part of the collaboration that happened. That's right. Including in the security model, right? And if we don't have that right from the beginning, then nothing else becomes possible. And there's a lot of that domain expertise within Accenture that helps us scale. One of the things that I've heard a lot today at the Accenture Executive Summit is this idea of thinking differently about failure. And this is an idea that's in Silicon Valley. Failed, failed better, fail happier, fail up, all these things. Fail fast, exactly. Fail forward. Fail team. But all of them. How do you, how, but how does a non-profit in New York City, how does it embrace that? I mean, as we've talked about a lot here just now, the people are the hardest part. That's a really different mindset and a really big change for an organization like Health First. But the business model of working with AWS too is pay as you go and everything. It's like failing cheap is very possible. We're not putting out huge upfront costs to turn something on. We can turn it on for pennies sometimes and do a use case. So it really does support experimentation. We've been, one of our successes I think is we really just, we try a lot of things. So we've had to learn how to do that and learn how to sort of either pull in more experienced people to help us or just cut it off kind of in some cases. So yeah, the cloud patterns and AWS business model just makes it really easy. And it's also key of course to have some quick wins that are highly visible. So it's my understanding that in the case of Health First there was, you know, whether it's reimbursement claims or there's potential fraud that can be detected that is a lot easier to start doing once you got your data into a common data lake and you've got world-class analytics tools that are available directly to the business analysts instead of requiring lots of hand-holding and passing data sets around. When you get those initial quick wins that builds the kind of enthusiasm that allows you to then take this from being a project that people are skeptical about to people really seeing the value. And people get excited about it too. So talk about some of the benefits that your members have seen from this. Sure, so again we have 1.4 million members. So just something pretty simple. Every health plan wants to prevent readmissions. So someone's been in the hospital and then they have to go right back with the same condition. That's bad for the member, bad for the plan, bad for everybody, right? So just being able to take a day's science model on our own data, train it up for predicting readmissions. Again, we have large care management, community, many nurses go out in the field every day and meet members, but now that we can give them a list of the 500 most important members and it's also self-service, it's in a dashboard that's running in Redshift and people can go and just get their lists. I mean, that's really profoundly satisfying and important to change our members' health outcomes. That's only one example, that was kind of the first model we built, but we have models for people being adhered to their medication. Just a lot of things that we can do targeted interventions instead of kind of having a bunch of business roles kind of in your head of who you think you should reach out to. This is the data is telling us who's most at risk. And sometimes empowering the call center personnel when you can give them access to data that allows them to really personalize that phone call experience with somebody, it's a relatively low cost way to surprise and delight the patient or the health plan member. And that then drives customer satisfaction scores, which are very important in the health care industry for all sorts of reasons related to accreditation, related to reimbursement and also frankly just related to enrollment and retention. I speak from experience when I say the best companies are the ones with the good call centers that you just are happy and you get off the phone and you don't want to slam it down, you're happy to talk to them. So final pieces of advice for companies that are trying to drive change through data analytics, what is the best practice, best piece of advice? Well, because you looked at me, I guess I'll go first. I'm sure we each have our answer to that. We always, it sounds obvious, but it's surprisingly often not the case once you get past the initial, five minutes of a conversation. Really stress, are we actually focused on a real problem as opposed to something that sounds cool or fun to go experiment with? Because these tools, as Ken said, these are, it's fun to play with these self-service AI tools, you can predict all sorts of things. Is it an actual pain point for either an internal customer or an external customer? Yeah, I think you hit it on the head as well. Best advice to starting this is get some wins, get some early wins, and then don't be afraid to experiment and don't be afraid to think outside the box. Shale? I think I'd say there are two pieces of advice. One is focused on strategy, like Dan was talking about before, because with tools like AWS where you can literally use your credit card to get started, you can lose sight of the big picture. So have a data strategy that is directly tied to your business strategy is very important. And the second is, instead of thinking about building a data pipeline for a specific use case, think about building a platform, a data platform that can serve the need of today and tomorrow as well in an architecture that is fit for purpose architecture, like Andy Jesse talked about today. So don't go for a Swiss Army knife approach. Go for fit for purpose platforms, products, models, if you will, that can allow you to build that platform that can serve the need of the future as well. Excellent, thank you so much. Shale, Ken, and Dan, thanks for coming on theCUBE. Thank you, thanks. Thank you. I'm Rebecca Knight, stay tuned for more of theCUBE's live coverage of the Accenture Executive Summit.