Skip to main content

Bayada’s Snowflake-to-Databricks Migration: Transforming Data for Speed & Efficiency

Summary

  • Snowflake to Databricks migration: This video showcases BAYADA's consolidation of Matillion plus Snowflake and SSIS plus SQL Server into a unified enterprise data platform on the Databricks Data and AI platform.
  • Medallion architecture implementation: Learn how the Databricks Data and AI platform's medallion architecture enables seamless data integration, advanced analytics and machine learning across critical business domains.
  • Operational benefits and efficiency: Discover how BAYADA projected to reduce processing times by 35%, improve reporting accuracy and cut reconciliation efforts by 40% on the Databricks Data and AI platform.

Bayada’s Snowflake-to-Databricks Migration: Transforming Data for Speed & Efficiency

Watch: Bayada’s Snowflake-to-Databricks Migration: Transforming Data for Speed & Efficiency
Talk by Elaine O'Neill, CDAO, BAYADA Home Health Care ; PradeepKumar jain Vimalraj, Sr. Director - HLS, Tredence Inc ; Venkatesh Guruprasad, Head of Data Architecture & Governance, BAYADA Home Health Care
Bayada is transforming its data ecosystem by consolidating Matillion+Snowflake and SSIS+SQL Server into a unified Enterprise Data Platform powered by Databricks. Using Databricks' Medallion architecture, this platform enables seamless data integration, advanced analytics and machine learning across critical domains like general ledger, recruitment and activity-based costing. Databricks was selected for its scalability, real-time analytics and ability to handle both structured and unstructured data, positioning Bayada for future growth. The migration aims to reduce data processing times by 35%, improve reporting accuracy and cut reconciliation efforts by 40%. Operational costs are projected to decrease by 20%, while real-time analytics is expected to boost efficiency by 15%. Join this video to learn how Bayada is leveraging Databricks to build a high-performance data platform that accelerates insights, drives efficiency and fosters innovation organization-wide.
Here’s more to explore:
Production ready data pipelines for analytics and AI: https://www.databricks.com/solutions/data-engineering
The Big Book of Data Engineering: https://www.databricks.com/resources/ebook/big-book-data-engineering-2nd-edition
See all the product announcements from Data + AI Summit: https://www.databricks.com/events/dataaisummit-2025-announcements

Chapters

FAQs

Why should I migrate from Snowflake to Databricks?

This video shows how BAYADA achieved 35% faster data processing, 40% reduction in reconciliation efforts and 20% operational cost decrease by migrating to the Databricks Data and AI platform. Additional benefits include real-time analytics and unified data management.

What is the medallion architecture and how does it work?

The medallion architecture organizes data into bronze, silver and gold layers on the Databricks Data and AI platform, enabling progressive data refinement. This video demonstrates how BAYADA used medallion architecture for seamless integration across finance, recruitment and costing domains.

Can I migrate multiple legacy systems in one platform?

Yes, this video demonstrates BAYADA's successful consolidation of Matillion plus Snowflake alongside SSIS plus SQL Server into a unified Databricks Data and AI platform, simplifying operations and improving data governance.

What future capabilities does Databricks enable?

This video shows how the Databricks Data and AI platform positions organizations for AI-ready data strategies including agentic AI use cases. BAYADA is positioned for future growth with advanced analytics and machine learning capabilities.

Full transcript

0:00 Introduction

Thank you everyone for arriving on time. Uh this session will be about Bayetta Snowflake to data bricks migration and on stage we have our three speakers there. Uh we have Venat, and So I would hand over to you to start the session. Thank you. I will start. Hi everyone. Thank you for coming today. Um O' and I'm the chief data officer at Bayata Home Healthcare. Um I will go through the company and who they are. It's very missiondriven um caregiving organization and I've been fortunate to be part of this company for 16 years.

0:28 Presentation

My background is traditionally in accounting and finance. I usually introduce myself as a recovering CPA, but I do not identify with that young lady anymore. Um but my journey throughout Bayata has brought me all the way through operations into the finance world and into data and analytics which I think is a a good journey to be on. You see how all of the transactions um connect throughout the organization. I work alongside Venitesh and I will let him introduce himself.
Yes, thank you. Uh. I my journey into BIO is less than two years into the company. uh I was brought in to standardize uh data architecture, data engineering, data governance in the company from foundational level. I'm uh happy to uh be part of this uh journey with BA because uh we we are not only mission oriented but we are very disciplined in our approach on investments and technology choices iterative mindset and uh you know just the way how data bricks has built the platform from the ground up to uh implement data engineering and uh architecture at scale.
We are looking at the foundational capabilities first before we go in and build vertical use cases on top of it. Right? So I'm excited to hear. Thanks. Thank you, Wenites. Uh I lead the client partner team at Tredence. Uh part of my job is to work with and Wenites, listen to some of the challenges that they have. Uh but most importantly, uh we get to bring the solution to Baya, right? So uh probably very privileged to be in this role where some of the challenges that and Mikadesh are seeing is something across all the healthcare organizations are seeing as well.
So hopefully we can shed a little bit light on how we are solving uh the migration project and how we are going to enable BIA to be AI ready uh most importantly uh start empowering them with agentic AI use cases in the future. Cool. Okay. So, let me tell you a little bit about Bayata Healthcare. Um, not very well known on the West Coast, um, but definitely known on the East Coast. So, we just celebrated our 50th anniversary.

2:49 About Bayada

We came to LA. There was about 3,000 of us. We did it all up. We had a wonderful time. But, we've been 50 years providing care to people in their homes where care should be provided. And we'll go into some of our beliefs there. We are a nonprofit. that is not you know for an organization of our size that is not very popular in this space. It's a very fragmented type of industry. There's a lot of mon pops um that are in this space or there's a lot of hospital networks that are now in this space thanks to co uh we just are over two billion dollars of an organization um and still on a good growth trajectory of about eight and a half% yearonear um so still growing within um our markets.
We service 44,000 clients in their home every single week. We operate on weeks, months, and quarters. We are very diligent in getting our financials out, our operating metrics out every single week. So, we watch our measures. 44,000 clients from children all the way to hospice. The continuum of care um is covered. We employ about 33,000 caregivers and clinicians across the organization as well. So every single week we are generating payroll in-house for 33,000 caregivers and clinicians across two 347 states um uh offices I'm sorry in 23 states across the US.
We are also in India, Ireland uh and New Zealand. Um and we have about 10 joint venture partnerships. So we've connected with some of the major hospital networks in some of our more saturated markets like Philadelphia, New York, and connected um their at-ome care is delivered by Bayata. So that is our relationship with the hospital networks. Our we are 0% chance of being sold. Our founder Mark Bayata, he donated the company in 2017 when he retired.
He made his son uh the CEO, but he donated the company because he wanted to make sure that we be here for another hundred years. And that is our lasting legacy that we will be in this caregiving um Bayata home healthcare for the next 100 years. I'll go into some of our core principles and how we're we're kind of structured and founded. We truly truly believe that care provided in the home is best. uh postcoid most other organizations also believe that care provided in the home is best uh healing happens in the hospital true curing happens at home and we do it across the entire continuum of service so I'm going to start on the left and talk about pediatrics we bring home kids from the neonatal intensive care unit every single week we are partnered with CHOP we have many stories about bringing kids home for Christmas that will touch your heart.

5:16 Core Principles

Um, and we do it with pride. The complexity of these kids and the diagnoses that they are coming home with are becoming more and more complex as we've been in this business. We are now bringing them home with equipment. We have hiku home intensive care unit. So, we're able to provide higher acute care at home rather than be in the hospital. That saves money. That saves the payer money. the payer is very interested in getting this work done.
We graduate then to adult nursing. Adult nursing is our adult population who needs uh nursing almost around the clock but definitely on a continuous basis. In our home health population, there are Medicare patients. They're usually coming out of the hospital. They've had a a wound, a hip replacement, um and they have short amount of care for a period of time. Hospice is our end of life. We have spiritual care. We have social workers.
We have um a whole team that readies a family for hospice care. Assisted care programs is our main uh home health aid program. So think of the bath visit or the changing that you need for your mom or your grandparent. We deploy those caregivers into the home to help them be independent. Um I have my great-g grandandmother um and she doesn't want her mom helping her get changed. she would rather somebody else come in and help her get changed.
We provide that care. They feel independent and feel like they're not depending on their children for it. Assistive care is the exact same type of practice for behavioral health. For us, that's autism. So, we specialize in uh some autism centers across the country and engage with a lot of uh diagnosis with children. Habilitation is when those children graduate and they have developmental um delayments, we help them uh with that service.
So if you think of all of the types of services that can happen in all of the communities, whether it's in a facility or in a home, we can provide that care and we bring the caregivers and nurses nurses to do that. Perfect. Uh a little bit about Credence. Uh we are of course we are one of the strategic partners for BIA. Uh we are close to 3,000 person strong building boutique with scale uh specializing in healthcare use cases.

8:02 Credence

Uh hot of the press uh we were named uh the partner of the year five times in a row. Uh this is something I sort of start saying uh which we feel very proud of because uh one of the biggest differentiator that we bring to the table is we don't look at uh problems from a technical point of view. We try to understand the use case that the customer is trying and bring in the relevant accelerators to make it happen. Uh so that's a little bit about us and uh of course uh yesterday we also won the growth partner of the year which is something we are very proud of.
Congrats. Thank you. Okay, let's talk about some of the issues the business problems that we are faced with um within our organization. Uh we are highly performancedriven. We mine our data the way that we can so that we can um understand um how we are doing on a weekly's basis, how we're doing against the care model that we have developed um with all of these service lines that almost act like individual companies. How do you consolidate all of that together and show the big picture for the organization?

8:54 Business Problems

How do you govern all of that? We are very entrepreneurial organization. We get a lot of requests that come our way that says can't you just and give me from a data perspective and we have to be able to keep that entrepreneurial spirit moving or we won't end up in 25 states 26 states. Um we are like I said purposebuilt and nonforprofit. We have our Bayataway which is a wallet card that we live by every member of the organization and we truly believe in the in this system and the value system.
If you know anybody who works at Bayata, they all know about the Bayataway. It is true and testament and has been around for about 25 years. From a future forward perspective, we're on the cusp of innovation. I'm going to say because I've been here for 16 years that Bayata is usually a fast follower, not an innovator. They're not usually the first one to go to market with anything. They'll watch what everybody else is doing, learn a little bit, try a little bit, and then move um in in one direction.
But from a clinical perspective, it's getting so complex and we are losing caregivers and clinicians like you would not believe. We need to maintain the workforce. We need to provide reliable care. Um so we need to innovate and we need to step up and we need to figure out a way to manage an ecosystem of solutions. Now uh you know everybody's into uh documentation. Let's reduce the amount of documentation on our clinicians.
They won't give up that power. they need an amount of power in capturing that documentation. So we are committed to new ways of educating them, new ways of developing them and giving them new tools to first assist them before we can completely automate something. I'll quickly go through our history. We like to call this the Bayata eras. This was our ta Taylor Swift moment of 2024. Um, and it's really how we've moved from a data perspective and an architecture perspective from the founding years all the way to where we are.

11:11 Bayada History

And again, you know, most organizations are going to start on paper. We built a homegrown DOSs system called Bears. And believe it or not, Bears still exists in our ecosystem as a legacy platform. Um, and it's really hard to unhook out of something that's so customized. It's really, really difficult to do that. So we've been on this journey of trying to unhook our ourselves from bears. We've been introducing some mainstream platforms like home care home base.
In 2017 we replaced great planes with workday which added more um tech stack to what we were not really prepared for. And then in this transition years is when we have like a soup of organizations um in our infrastructure and we're not managing them like they should be. There's no coordination. Everything's happening in a silo. Homeare home base sits over here. Workday sits over here. How does everything communicate?
How do we bring all of the data out? And while the fingers on keyboard person who's trying to do a schedule or approve payroll, they can move forward. We're really not being smart with the data that we have. We're not coallesing it down into what does this mean for us? How do we do better? We're just grinding the grind and doing it over and over again. And that's not scalable, right? So we're transitioning, our strategy is actually transitioning into the scalable years and that's when data bricks entered into uh the Bayata relationship and the offerings of how do you streamline, how do you orchestrate again I will say this is an ecosystem of solutions how do you manage an ecosystem of solutions and how do you ensure that as you scale you're able to input more you're able to you know decommission more. you're able to keep that alive from a continuous improvement perspective.
This is not a set it and forget it. We set it and forget it 10 years ago when we put our data cubes in place and we thought we were ahead of the game. We had clinical data, we had financial data, but then we added all of these practice management systems and we lost control. So, um, thanks to Venites, data bricks was introduced uh to the organization to help us solve some of those problems. I can go into more of the problems that we were trying to uh accomplish.

13:43 Simplification

But again, from like a a structure perspective, where where are we looking to go with it? Simplification. Simplification is the name of the game. Well orchestrated simplification ecosystem. These are buzzwords for us in the organization. So we need to be able to ingest all of the data. Benitesh will tell me we have 65 plus sources of data that are coming in at any given time and we need to be able to bring them in in a standard format.
We need to take down the silos between all of that data and be able to put it together. Our patients are our patients are say are served by multiple practices. You can be on adult nursing and receive habilitation services. You should not look like two clients to us. you should look like one and I should be able to see a patient 360 on that particular client. Improving clinical care. We're on we have to we must we are losing the caregivers through burnout um because of it.
Uh the federal government is introducing even more regulations around how you perform care and how we'll give you a star rating. You're a four star, a three star, a one star based off of how they think they will measure you. We have to stay on the cutting front of how they are implementing measures so that we're able to keep up with them. A lot of the data that we do today is manual. There's a download from one, there's a cube extract from another, there's a report over here, and our analysts combine it all together in their own data set.
And then they turn out the data and they build it and they build it. A lot of manual effort, a lot of time means it takes a long time to make decisions in this organization and the decisions to make positive impact are really difficult to make. That is where we are and those are the solutions that we're looking for. So as we were introduced to um data bricks, data bricks then introduced us to tradance who also has some ways to be able to accelerate where we are going.
We are already behind the eightball here. Want to add? Absolutely. Uh next slide I think. Sure. Uh Price, do you want to just share a little bit about uh some of the challenges that you were facing right? Uh not just from a technology friend perspective but uh what were some of the things that you had to think about when we were in snowflake and you know what made you decide to move into data bricks? Yeah. Yeah. Thank you.

15:55 Challenges

Uh so I think the left hand side uh summarizes the challenges. I think most of you in your organization you probably have similar things uh you know that we have listed there. Uh you know we when we have seven to eight different uh practices we whether we like it or not the origins of it becomes a data siloed environment. Data is all over the place. Now we need to figure out what to do with it to make the best decision possible. uh because datas are siloed you can't make you know insights better to the end user more actionable and you know more uh you know specific to their needs and uh we have multiple data platforms so we had SQL server data warehouse we started our journey to snowflake before I joined now we retiring both of them to into data bricks right so we went from three data platforms into one so multiple platforms creates uh uh challenges on the scalability of the whole data process as a whole and uh we we as a organization uh we are entrepreneur you know as bound in terms of how our offices run.
We wanted to gain them give them access to the tools they need so they can make their own decision. So it doesn't need to be a central data team giving them the reports or the insights. They want we want them to be self self-service driven in how they manage their own business. Right? So that was a challenge and when you put all these things together the big elephant in the room is the governance because it's not about technology it is process business process people all of them combined with what makes a a governance actually work right so what we at a strategic level what we decided was go to the right hand side there we want to get to a unified data foundation so collapse all the multiple data platforms into one and that's where our strategic investment in data bricks came in. it is operating our acting our enter uh you know enterprise data lake and also our reporting analytic engine sit on all all together in one and the second one thing is very important I think we we highlighted the insights inefficient insights on the left hand side to real-time decision insights uh back 10 15 years ago I've been in the data lake data analytic domain a lot and uh 15 years ago the biggest uh you know uh selling point for organizations for when you know for platforms was uh you want to be um uh you want to be actionable insightdriven company right so hey you want I will give you actionable insights right I think we live in an era where action limits site will not move the needle we need to get to decision based insights where you are not telling the officers to say hey your recruiting is down because your job description of the portal is bad but you want be able to move that needle and say hey can I build an agent where it automatically update to job description for the next month so your recruiting catches up again right so I think that's uh that's the innovation that we are you exploring and I you know we uh recommend that for a lot of you guys who are you know thinking about insights now don't stop at actionable insights talked about decisions you can take and give that capability to your end users how they can actually make those choices with some automation built into it right and a background and because now we are able to collect all the infrastructure into one place.
Now we created a you know uh highly scalable you know platform to meet uh our our needs and also move the needle of data consumption to the self-service analytics and later in the slides we are talking about some tools that we are you know rolling out to our end users and innovations that we are building around it uh to make this data you know accessible to our end users in the moment they need to see this uh so they're not coming to a data team for building a report or building insight uh you know uh And obviously the governance piece is the foundational thing that we are starting the journey in the company where we are trying to build the right governance framework to make all this happen as we move forward.
Right. Something that this slide does not talk about is uh making biata an AI ready company. So Ela as a CDO right uh how do you I mean these are some of the tactical challenge that challenges that you see but how did that AI readiness play a big role for you to do a migration towards data bricks and what were some of those cultural challenges that you saw uh making the transition as a co Sure um I think there are a couple of challenges one is that we are a highly tenured organization so it is really hard for people to change.
When you're a highly tenured organization, they get set in their ways, their workarounds. But when you bring in people and you observe how folks are doing their work and you ask why, why are you doing that? Just because there's no other way, this is broken. It doesn't work. Um, you then light a little bit of a spark and a little bit of a fire of people saying, why am I doing this all the time? um covid was a a big wakeup call I will say for the organization as well on what's important what's not important what should we doing how do we become more productive how do we have better utilization like these you know operational effectiveness um and efficiency plays started becoming top of mind and when you go into the scale years those are the conversations that you have how do we do more with what we have today how do we you know work smarter rather than uh uh rather than harder.
Um, and you know, we had conscious conversations around them, but the demands are coming in from the business. The demands are coming in from the clinician. The ideas are coming in from the clinician in the home that's saying, "Hey, I have this app and it can record my notes. Can I use it?" We're saying, "No, it's not within the firewall and there's no security. You can't." But everybody is innovating and ideiating in some way, shape, or form.
And as an organization, we're taking a step back saying there is a part of the organization that is change ready now and there are the high adopters and why don't we use case a certain number of solutions and go about it that way. It doesn't have to be this massive change that you communicate to the organization. You don't have to scare anybody. There's nothing to be afraid of here. When you start getting that like small amount of adoption and implementation and folks don't even realize that it's happening, things just feel simpler and seem simpler.
You don't have to. It's not this big hairy thing, right, that I need to be afraid of. It's not a monster under your bed. It is welcomed when you say to folks, hey, you've been working with an agent the whole time. Makes sense. And uh probably the next slide is going to be a little technical heavy. So for the technical folks out there uh they'll probably understand this. Uh winkesh can you share I mean this is the legacy architecture of VA right this is post pre data bricks era right uh can you share a little bit about your legacy architecture what were the bottlenecks that you were facing and most importantly what triggered this change uh for you yeah so I think uh we have put notes under each block I think that kind of summarizes the challenges that we were going through uh as I mentioned earlier we have three different data platforms So you know so that results in three different versions of sources of truth for our business right which is really hard right finance teams looked at our SQL warehouse reports and they think that's the source of truth and we started journey into snowflake warehouse and you know our operation folks said hey this is my source of truth now you create the rift right you know whose number is right or whose number is wrong right so and uh on the right hand side it's about 2D reporting experience right which which is how the industry evolved, right?

23:07 Legacy Architecture

VI was evolved as a 2D reporting uh you know uh experience and in still you know for large organization is still a 2D reporting experience right so either they zone in on a tool or zone in on a platform and build a whole BI stack around it right and you build all these smarts and cubes and fact tables you know you throw all kinds of technical stuff on top of it and make it so big at some point that won't scale both operationally and financially also right as as companies grow.
So those were the three main areas right to say hey we need to take a pause and really see when you go to the scale area scale years right what do we have to do that needs to be right to the organization and when we need to go from 2D BI you know reporting to a self-service analytics is a major leap right and you cannot have three different platforms and three different versions of source of truth to make that happen that won't work right yeah well I mean we we just saw the challenges that we are facing Right?
But then you know we live in a world where there are so many options when it comes to data platform. Right? Uh you know I can at least name 10 of them in my sleep. Uh but when there are just so many options right what are some of the key KPIs or some of the key criterias that you take into consideration not just from a use case perspective but from an enterprise perspective? what are those evaluation criterias that you had gone through uh to make the decision to migrate towards data bricks?

25:13 Key Criteria

Yeah. So uh you know again I you know we are in a data bricks summit. I don't want to say snowflake is very bad. So we have to move to data bricks or vice versa. So uh but uh each one each of these platforms comes out its own strengths and weaknesses right and most of these uh you know our our view of these platforms is also time bound right because all these companies also evolve in features and capabilities over time right uh but our biggest uh uh you know uh driver into data bricks I would say is based on four different key factors right if you look at from a technology angle is one is we don't want to uh get into a lockin solution with any vendor, right?
Data bricks allows that they are not a locked in platform, right? It's all built on open source. Uh the second thing is scalability and performance and even with the workloads that we were running on snowflake now that we migrated to data bricks sideby-side workloads you know we are not like a you know pabyte you know you know data company where we have a lot of data pabytes like maybe like a banking financial insurance companies have but uh even at the scale that we have we have probably like you know close to 30 terabyte you know enterprise data but even in that scale you know data bricks outperformed same piece of code 30% better right you know compared to snowflake right both on cost and you know performance as a whole right like by like like by like comparison right and uh when we were moving into the machine learning AI agents and all those stuff data bricks had a leap uh you know forward in terms of capabilities and when we looked at uh where snowflake was you know two years ago and where datab bricks is there even when you compare feature functionality very basic stuff that we on the tech side is say can this be integrated to GitHub right can the CI/CD pipelines can be built uh snowflake was very early in the journey and two years ago probably they caught up to it but data bricks was like two years ahead of them right so those were all the key technical factors right but I think the from a business perspective is also important because when you make this multi-million dollar you know multi-year journey more than the money it's all about right Can we reach do we reach a point where it'll be point of no return or should we can we figure out what do you make out of this investment right I think that's where the being open source you know uh being able to scale you know control your own destiny really made a big role right and we also looked at other tech tools right on the right hand side like we have AWS native tools like you know uh we are AWS shop uh we had data robot in house right we we did some PC's with it that it really not did not solve the entire data journey.
It was only isolated to building machine learning models and you know those were the factors that was primarily used right to you know decide that hey data bricks is the right strategy for us. I remember our CEO coming into me when I had that contract in front of me with data bricks. Um, and he said to me, can you promise me that we will not be replacing this three years from now? And I said, I promise you we will be building on it three years from now and three years after that and three years after that.

28:39 Why Databricks

So it was not a decision that was taken in light. And um I think Venitates started off by saying um we have become really stringent in our investments and where we make our investments as an organization. We are funded by Medicare and Medicaid. We don't have room for lots of mistakes. Um so this was a strategic bet I will say at the time. Um but one that we don't look back on with regret in any way, shape or form. Perfect.
Uh next slide. a little bit about how credence achieved uh with the migration right. Um as I said uh earlier uh one of the key differentiators that we really bring to the table is the use of our accelerators right uh now if you sort of take any of the migration projects that are out there without any accelerator I would easily say it's around 9 to 12 months right every organization takes the journey of 9 to 12 months to migrate from one data platform to the other but with some of the accelators that you see over here which is T- Injust Tiscover TDQuest we have these pre-built notebooks, pre-built knowledge that we bring to the table that actually accelerates the migration and uh with Bayata we have managed to do that entire probably not entire I know Ela will ask for a project plan not yet but uh we have managed to uh do a lot of these with the accelerators within the five months of our engagement so it's really speed to value and most importantly uh these are proven accelerators that these are not like you know something newly built but these are tried and tested with our other customers evolved over a period of time and that can really add value towards the migration journey that we bring in.

30:47 Key Use Cases

Uh now what we earlier saw was more from a technical perspective but a little bit about what we do with the data is uh some of the accelators that you see over here is what we call as empowering a cal delivery through AI. So many of the organizations that we work with especially the payer providers some of the common challenge that challenges that we start seeing is they have challenges with their EMR data claims data eligibility data different file formats and so on and so forth.
What really come brings to the table is our accelerators are pre-built datab bricks native solutions that can take all these data across different data sources be it claims clinical and so on and towards the right are some of the key use cases that we empower our customers with uh some of the common and I get this a lot because I'm in front of the customers day in day out I get this a lot some of the challenges the customers really say is pra uh we have challenges with our enterprise master patient index or data mastering we have challenges creating remember 360 which is a com combination of different data sources to create the golden record uh that's where accelerator comes to the picture so these are pre-built uh tested and tried proven across different customers evolved over a period of time enabling the users to achieve their use cases faster but most importantly accurate and accurate is the KPI that many of our customers evaluate us with and that's what our accelerators basically bring to the table.
Yeah. Now uh with our accelerator what you earlier saw was the legacy architecture of Bayata right uh they were using snowflake SQL and so on. Uh what you see in the slide right now is the current future state architecture that we are heading towards. So uh you know let me start with right uh when you see this architecture of course there's a lot of uh data bricks icon but uh what thoughts goes into your mind when you implement such a future state architecture and make biata AI ready for 2026.

32:20 Future State Architecture

Sure. So um I'll repeat myself and say orchestration and ecosystem are very top words here. It can seem very complicated when you're not in the tech space and you don't know how they all connect with each other, but it is easy for me to show the business the value that they are getting from an architecture that looks like this. The seamlessness, the way that um the relationships are built between each of the tables. Um folks are learning how to model data for the first time in in forever from a business perspective.
So, we're readying the organization for that. Um and also seeing like we call this the chassis, right? We're building the chassis right now and the chassis for us is um like a Toyota. The Highlander and the Lexus is the same chassis. We're not at Lexus yet. We're at Highlander. And you can put all of the bells and whistles, but I will guarantee the organization that we have the chassis in place in order to be able to do that and to power that Lexus to the next stage.
So, that is our focus. Um and that's where we're grounded on and the organization can get behind that um and buy in. And Wenardesh uh just switching to a technical perspective right uh can you share a little bit about your insight what went through the thought process on this future state and a little bit about the different accelerators that we have been using. Yeah. Uh so as uh mentioned so they have accelerators that can address different uh steps in the data flow journey right so if you want to do data quality checks if you want to do ingestions if you want to create uh data models right uh and you want to enable the semantic layer that can go into self-service consumption they built all this accelerator at every step in the journey and also underneath all of it is the code conversion so it can autom automatically read code from one platform and convert the data bricks equivalent right so I think combination of all these things is what made our journey faster right you know and uh meet where we need to go right in terms of a future grow so uh what's the future of VA looks like with this new architecture so from my perspective and you know when I when I go on the show to talk about what where I think we'll be I think every director in every office will be able to glean their own actionable and decision-making um insights at the tip of their fingers.

35:01 Future of VA

Right? Natural language is allowing you to remove your dependency from the analyst. Give me a list. Who are the people? What should I do? Why am I off? Um for a a director in an office who runs a team and services a community to be able to quickly action on those items. We're weekly business. We look at our data on a weekly business basis so they can be that tactical tactical um in their approach. So I see them raising their engagement scores.
I see them raising their ability to impact their communities to be their ability to see um uh good clinical outcomes and how that relates to a payer contract and bringing both of those together and not them sitting in isolation. So my focus point sits on at the director of every single office because they're that's the role in our organization that drives our business forward. They're the ones who engage the caregivers and clinicians.
They're the ones to say, "Hey, I don't need to be in the home. I can do remote patient monitoring. Help me do that." They're the ones to say, "I need to optimize my job post performances and I need feedback on those job posts and I need to be able to implement changes." Bring in an agent to do that. I don't want to do it. We all know that it's not um getting the hits it should be. Change it. Change the keywords. Do some search engine optimization.
Give them assistance to help them in decision-m. Give them assistance to help them manage. Give them assistance to help them um be more compliant. If you saw compliance in home healthcare, it would make your toes curl. It is labor intensive. It is paper driven. Medicare still uses fax machines. Like that is the reality that we live in. And so we are trying to innovate and move forward this type of an industry so that we can and will say this too, we can kill the fax machine um so that we're not sending faxes all the time.
So when we level up, I will say the directors in every single office, we we are dynamic um in the industry and you change care for the entire industry which then lowers the cost of delivering care to the entire continuum. It is cheaper to receive care in your home than it is in the hospital and we keep people out of the hospitals. And uh any future use cases uh that you're going to focus on? It's funny that this says future.

37:37 Future Use Cases

This is actually now and this is not all that much future. Um but we do you did talk about like mastering being a problem. Um it is a problem for us as well of which uh Tredence is helping us solve. Um but you know all of our entities across the organization mastering that makes tremendous help in just generating payroll and revenue and claims right. Um marketing and workforce analytics has been an area that we've been focused on and have not moved the needle um for the past three and a half years.
We need to be a lot more thoughtful. We need to deploy agents in this area because it's repetitive tasks that we do. Send out conditional offer, send out pre-boarding paperwork. Like you don't need to put your your talent acquisition partners behind that. Um we need to relay our valuebased agreements which the payers are coming to us with the table with the outcomes that we're achieving on those. Um so the director is responsible for the outcomes and the enterprise is negotiating the contract.
They need to come together better and our data is going to help us do that which means we canre increase our reimbursement which means we can pay a nurse more and then the cycle continues and continues. Um AI documentation for clinicians um is definitely an area that we've dabbled in. We've not like hit the nail on the head there. Um some organizations just do a piece of documentation and not the whole workflow and you really need to see it end to end.
Um so we continue to u make some dents in that. Um and then our clinical QA process which is essentially auditing. When the state comes into your office and they want to see your charts, we need to be able to automate that. We want them to do their auditing remote. We don't want them in our offices doing that auditing which rem means we have to remove all of the paperwork from that process and we are definitely hellbent on making that happen too.
But these are now I will say I will come up with another list for future. Yeah. Okay. Yeah. That that means I need to update those slides. Uh well we still have a minute but uh why don't we uh happy to take a question or two from the audience.

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.