Healthcare Finance Transformation: Unity Catalog and Genie for Data-Driven Decisions

Summary

EY's finance transformation practice and a major healthcare organization show how unifying cost, reimbursement, claims, and service line data in Databricks Delta Lake with governance through Unity Catalog cuts finance close cycles and enables confident enterprise decision-making.
A critical insight from this video is that 80% of finance data sits outside finance systems, requiring finance teams to own data definitions and quality rules rather than passively waiting for IT to deliver governed datasets.
Genie enables self-service access to governed financial data through natural language queries, while a medallion architecture with a finance-specific semantic layer eliminates manual calculation inconsistency and reduces analyst workload across the organization.

Healthcare Finance Transformation: Unity Catalog and Genie for Data-Driven Decisions

Healthcare finance leaders need faster access to trustworthy data across clinical and operational domains. this video shows how a major healthcare organization transformed finance operations using Databricks Lakehouse Platform, moving from siloed cost and claims data to a unified, governed approach that cut close cycles and enabled confident enterprise decision-making. By unifying cost, reimbursement, claims, and service line data in Delta Lake, governing definitions and quality with Unity Catalog, and enabling exploration through Genie, finance gained the foundation for AI-ready insights.
The transformation required intentional strategy across organizational, process, and technology dimensions. Key elements include data product owners ensuring quality at the source, finance-owned governance and quality rules, medallion architecture with a finance-specific semantic layer, and Genie for self-service access. Results include clearer variance analysis driven by business and operational metrics, consistent definitions across the enterprise, and significantly reduced analyst workload through semantic models that eliminate manual calculation inconsistency.
🤝

Chapters

FAQs

How did a healthcare organization transform its finance operations using Databricks?

The organization unified cost, reimbursement, claims, and service line data in Delta Lake and applied finance-owned governance rules and quality definitions through Unity Catalog, replacing siloed and manually reconciled data sources. This created a consistent semantic layer that reduced analyst workload, enabled clearer variance analysis, and provided the data foundation needed for self-service AI-powered insights.

What does it mean that 80% of finance data sits outside of finance systems?

According to this video, 70% of CFOs report that data blocks timely insights, and 80% of the data finance teams need actually lives in systems owned by other departments such as operations, clinical systems, and supply chain. This creates the gap between AI ambition and data reality that finance transformation must address by having finance actively govern and own data definitions across the enterprise.

What is Genie and how is it used in healthcare finance?

Genie is a Databricks capability that enables business users to query governed data through natural language without writing SQL. In the healthcare finance context described in this video, Genie provides self-service access to the unified finance data layer, allowing analysts and finance leaders to explore cost and reimbursement data directly without routing every question through engineering.

What are the three transformation levers for healthcare finance AI readiness?

This video identifies data products, automation, and value orchestration as the three transformation levers for modernizing finance operations. Building governed data products at the source is positioned as the essential foundation, because automation and AI-powered analytics cannot deliver reliable or trustworthy results without consistent, well-defined data to work from.

Full transcript

[00:08] Good evening everyone. It's such a pleasure to be here. So as it is required for all sessions, I want to draw your draw your attention to the disclaimer that's on the slide here. Take a moment to read through that and then we're going to breeze through the the next slide where it talks about completing your survey. Um so if you
[00:25] have attended similar sessions, if you want to go to the next slide. Um this is just uh instructions for you to make sure that you are able to complete the survey at the end of this session. So with that I want to jump in. Thank you everyone for coming. It's such a
[00:41] pleasure to be on the stage here with my colleagues. Um I want to start off by saying as I think about this topic the future of finance and my role at EY I lead finance transformation strategy and vision for the Americas at EY. I spend a
[00:57] lot of time with CFOs talking about the questions, the challenges that distress them on a day-to-day basis. And primarily that deals with how do they make decisions better and faster. Now, that's easier said than done. And if you put it into perspective,
[01:15] um we want to make sure that we talk about how they answer for some of these challenges. If we want to kind of look at the slide here, I want to call out for you a quick reality check. By show of hands, how many of you think that your organization is AI ready
[01:31] today? If you want to put up your hands, one, you are using AI to drive decisions actively today. Oh, two hands, three hands. You are pivoting towards using AI, but you haven't figured out how to scale AI as
[01:47] of today. Oh, that's great. Almost 80%. And how about the ones that are still trying to figure out how to interpret the data and what the performance was for the last quarter and the last year? Okay, I got some laughs. So, I
[02:03] appreciate that. And so, that really the last two things is really the reality of most of the organizations we work with today. It is that gap that we see between AI ambition and the data reality. And if you think about that for a minute, the two stats that here are
[02:19] very telling, right? 70% of the CFOs say that the data blocks timely insights. So that should not be a surprise. What is more surprising is the stat that follows. 80% of that finance data sits outside of finance. And that is the
[02:36] reality which creates that disconnect we talked about and the the raising of the hand points towards. So again, if I take that a little bit forward, it should not be a surprise that data hence is the new center of gravity. I'm sure over the
[02:52] days that you've been here in the conference, you've heard that loud and clear. Data is the one that matters. So clearly the slide talks about data being in the center of gravity. But what I want to point your attention to is what the picture is changing from a finance opodel perspective is how do you
[03:09] reimagine that operating model? You start with the left hand side where the typical transaction owners, the process owners are now evolving into data product owners. They own the domains, the revenues, the claims. In this case, I work a lot in the health
[03:25] and life sciences sector cost and pricing and they are generating data. And the goal here is not simply to generate the data but to generate the data in a way such that it is correct at the source to begin with. So you're not spending all this time as you're automating the process at the back end
[03:41] reconciling and coming up with manual interventions. That sounds good on paper, but if you take that one step forward and you look at the center of this chart which talks about governance and this is where finance's role has
[03:56] been changing dramatically for organizations which have been able to do this well is it's all well and good where you're able to describe the data but if you don't put the governance at the center of the equation where finance owns the definitions, the quality and
[04:12] the vigilance over the data that's when this operating model starts to fall apart. So that's your foundation that connects the data creators with the governance. So now if you have the foundation right, you start to look at the right side of that equation where
[04:28] you can clearly drive insights from a value perspective, but you're looking at the value and the compliance functions coming together and that's where the real time decision- making becomes a reality. So if in order to bring this whole picture together, the one big takeaway I
[04:44] want to leave you with is you're making sure that your traditional transaction process owners are becoming those data product owners that are making sure the data is coming in right first because AI can only scale with better data. Bad data equal to bad results. No surprise
[05:01] here for the audience. Then you want to drive hyper automation which is making sure that you're moving towards more autonomous but more touchless processing with specific exception handling. And then the last lever for that is okay with all that said you're then enabling
[05:17] finance to be the value orchestrator to really drive decisions with the insights in hand. So one may pause and say well all well and good it looks good on a slide. Does it really work? And so to answer that question, we
[05:33] wanted to share a real life example with Shivang and Rashmi of the work we've done at Humanana and how we've brought this future um and made this a reality. So I'll turn it over to you Shivang.
[05:51] So uh starting with introduction uh Shivang Patel uh been u I'm a director for data strategy innovation and AI at uh Humana for finance uh been in the healthcare industry from a payer side for around 18 years and uh been in the financial system space for around 12
[06:08] years um and background I have a specialization in computer science and a specialization in finance. So I've been in the cusp space uh for a very long time and u I'd like to start with yes we talk about we talk a lot about
[06:23] technology right uh I do want to ground us I think one of the first things before we even look at the slide and we look at the numbers on there is the intent to make intentional investment in a finance data strategy I think that's
[06:40] the key and yes technology is a big part of But if you have an intentional finance data strategy that helps you drive a lot of it. It helps you drive your architecture. It helps you drive your organization. It helps you drive your data. It helps you drive
[06:57] everything, right? Uh everything that I'm going to talk about. So first and foremost, what we did is we invested intentfully, we invested people, we invested, you know, resources to set that data strategy. and uh my
[07:14] partner here from EY she'll go into the details of it and I'm going to ground us in the foundations on on that on what that strategy stands on right so diving into it on your screen you have number one and number two that talks about
[07:31] that's not as much technology that's more about process that's more about governance that more about organizational structure and processes right three four and five uh I'm going to use BA's you know raise of hand how many of you heard the word context today in different keynotes or mean anyone
[07:47] anyone burn I'm kidding so that kind of forms somewhat your context right a focus on context and then going into as a result we get a AI ready finance first data strategy going into that right so I'm going to talk about that so starting
[08:03] with that once you have um a data strategy the first and foremost is your finance uh data data governance, right? So, data governance, we all do it today, but we all do it with
[08:18] what we have, the technology stack we have or the ways we have today or what we have done in the past. This is talking about doing it and we have heard this in the keynotes today. It's with doing it with AI in mind, with context in mind, in different technologies in mind. So, going forward, we have to
[08:35] think about data governance um differently. So we can enhance or we can create and holistically end to end right. Second is unified u architecture. So the data strategy gives you an architecture u I'm not going to get into the architecture but I based on that
[08:53] strategy in my role I intentfully look at all the systems and design holistically going in because that's how you can move your strategy and you can move the needle on your strategy to get to that number six. We all want to get to number six but we can't skip. we have
[09:08] to take all the other numbers together uh to get there right so that's the key um so once you have that holistic I'm going to use the word holistic complete you know use whatever you want you have to do end to end um that's when the
[09:24] number three four and five come in talk about master data reference data management u if you if there are people here from finance background I'm sure you don't do any vookups in Excel or any have no reference reference data, right? Um, but sarcasm aside, it's very
[09:42] important to manage your reference data, especially with with finance in context. That's going to lead into your controls, that's going to lead into the quality of the data you need for finance. Um, so master data, reference data, intentful technology selections, intentful part of
[09:59] your architecture, design and governance, right? So then going into metadata, DQ and controls, it's it's not necessarily building over it. It's obviously complimentary. The more richer metadata you have, metadata equal to any
[10:14] guesses context, right? So it's part of it. U so it's going to help you build that more data quality rules based on data quality engines that you get in any modern platform today. And then controls, you cannot do finance without controls. You have context. I can
[10:31] guarantee you it's going to help you build your controls in the future much better and much more robust. Uh going into unified semantic model, you've done all the hard work and now you can now derive the benefits if you have a controlled intentful semantic model. You
[10:48] can increase your reusability. You can increase your quality because you'll have the same numbers. But you have to be intentful. That's where the one and two come in to help you make number five better, right? And then lastly, if you do everything right and intentfully, you're going to build a lot of context
[11:04] over a period of time. And hence, you'll have a AI ready speed to market. All the good things that we dream about are going to happen if you do one through five and get to number six. So that's uh I think without technology, but those
[11:20] are the foundations. technology all exists but sometimes you forget the intent in creating the structure um which will drive your transformation which will drive your finance forward. So with that I'll hand it over to Rashmi who will take us further.
[11:36] Thanks Siobhan. Um hi everyone my name is Rashmi. I'm a senior manager at EY in the finance tech consulting practice. I know it's me and my last slide sitting between you and your happy hour. So I will only take five minutes. Um and just
[11:51] quickly cover this right so you all heard about strategy it's still a paper exercise like we defined everything on paper now we have taken it further at humanana and starting to execute it so I'm going to just break it down right so
[12:08] you have the strategy now how do you go about execution so I'm going to just break it down how are we doing it what are we doing it and why are we doing it so when it comes to how we have been very intentful. So you could just like if you start you know going through your
[12:24] data transformation journey we all know it's going to take you years and years. So when I say intentful we wanted to prioritize the use cases that are of significance to finance that drive value that drive business outcome. So we have
[12:41] identified a few use cases of course we work in the CFO org on what do you want to see and I'll talk about those use cases in a minute but then where is data bricks in all of this right so at the core of this data transformation is a modern data platform and we are using
[12:58] data bricks for that so data bricks is helping us get all the data in one place standardize it and make it usable for finance shang talked about metadata data we are using unity catalog to make sure we keep
[13:13] the descriptions the definitions bna talked about it finance has to drive those definition so we are tackling that it's a um shang talked about architecture it's a medallion architecture you may have heard about it you may already be implementing it that
[13:29] is what we have chosen at humanana and we always keep coming back to that target state architecture so we can govern not just these use cases but everything else because you're not going to stop the work that is going on once you have the strategy in place but once you have the guard guard rails in place
[13:45] that is what we are using the architecture for so for these use cases medallion architecture what the one thing I'm really excited about is semantic layer right we all heard semantic models and why I'm excited about it because we are not doing the traditional BI semantic models we are
[14:03] building the finance semantic layer in data bricks so what is it going to do and I don't know how Many of you have built a semantic layer or are familiar with it. We heard about it in last two days a lot. Right? So what are we trying to do with it? Finance like BA said 80%
[14:20] of the data sets outside but all the calculations that are finance specific we defining them in the semantic layer. So not like everybody who needs to report a KPI to whoever they need to report it to have to calculate it by themselves because not two calculations
[14:35] are going to be alike and then you run into the reconciliation nightmare. So I'm really excited about the semantic layer finance specific semantic layer that we are building at Humanana and we are using data bricks for it. Um on top of that semantic layer we have um
[14:51] PowerBI dashboards because that's not going away. Your traditional reporting is there but we have also started putting Genie on top right. So we have set up some foundational capabilities there and like our vision is we are going to just expand it. So that's the how part right the architecture what's
[15:07] the approach is the use case based approach we want to drive some outcome. Now what are those use cases right? So for our CFO uh they don't want to be reactive they want to be proactive. So we are the first use case we are tackling is called a driverbased
[15:22] reporting. What's that that's going to enable? It's going to give them visibility into the financial and operational drivers that are behind the business performance. So we as data people we like to talk about drill downs. So wi with that intention they
[15:38] can drill down and really take thoughtful actions on what's driving this variance rather than waiting for that annual report and say oh I missed my targets or oh I missed my budgets or this forecast was not right. So that is what we are trying to tackle when I say
[15:54] driver-based reporting. Other things since you know it's Hima healthcare we are also looking into claims analytics and there are many more use cases that you know we're trying to drive value. So think of it as I'm not going to fix like you know get all the data into the data
[16:10] layer and then figure out what we want to do with it but let's drive with intention as as I call it the right hand side of my architecture. Now why are we doing it like of course our vision um we are not there where we want to be right so there is always an intim state before you get to the target
[16:27] state so our vision of course is um eventually we want all our finance leadership to have these uh the data at tips of their hands so they don't have to wait for analysts to answer a simple question that takes weeks right like that's the goal I think we all want that
[16:44] target state and we all vision for it um so that's our goal That's where we are going with with finance like fully self-service and not just like you know genie but the advanced um you know technology capabilities we heard today
[16:59] and I was telling the team like what I'm most excited about like when I talk about semantic layer and driving value is the announcement they made about genie ontology I think that's going to be great and we are really looking forward to bringing that in but I think we've taken on this journey it's it's
[17:15] going to be a multi-year journey I think we all understand that and there's no kidding about it. Uh but that's the intent right we've been very intentional about it and um that's where we just wanted to share our journey and if you have been on a data transformation journey be it for finance and I guess
[17:31] like one last shout out I'll give to Hima is um finance is really driving transformation right so they are really pushing the enterprise to think about um some of the things that you talked about Shiobhan right like metadata reference data finance is really spearheading all
[17:48] of that so I think it's we are very proud of the work we are doing and we just want to share with you guys. Um if you have any questions or want to share any of your experiences please you know let us know and we'll be happy to hear it. Um but I think we are at our almost
[18:03] at 20 minute mark. So I hope you found this useful and um yeah one more minute and then happy hour. So, all right. Any questions?

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.