Sponsored by: Qlik | Unlocking Service Efficiency: How Healthcare Uses Databricks and Qlik

Summary

This video shows how Qlik and Databricks integrate to deliver healthcare analytics: Databricks provides the governed data and AI layer while Qlik surfaces near-real-time visualizations and a service engineer AI assistant for equipment troubleshooting.
The integration architecture uses push-down SQL from Qlik into Databricks, MCP server and Genie integration for natural-language queries, and Unity Catalog for governance and lineage across structured records, images, and diagnostic data.
A live demo shows a service engineer entering an error code and machine type to receive targeted troubleshooting guidance—including relevant images and videos—dramatically reducing equipment service times and improving patient care outcomes.

Chapters

FAQs

How do Qlik and Databricks work together for healthcare analytics?

The integration uses push-down SQL from Qlik into Databricks so that heavy data processing stays on the Databricks Data and AI platform while Qlik handles visualization and user-facing analytics. This video shows how MCP servers and Genie integration extend the architecture to support natural-language queries alongside traditional BI dashboards.

What is the service engineer AI assistant demonstrated in this video?

The service engineer AI assistant lets engineers enter an equipment error code and machine type to immediately receive targeted troubleshooting guidance, including relevant images and instructional videos sourced from data stored in Databricks. This video shows the assistant dramatically reducing the time engineers spend diagnosing equipment failures, which translates directly to faster patient care.

How does Unity Catalog support governed healthcare data in the Qlik and Databricks integration?

Unity Catalog provides lineage tracking, access control, and governance over the healthcare data assets—structured records, images, and diagnostic data—that flow through the Databricks Data and AI platform into Qlik. This video covers Unity Catalog's role as the governance and AI asset layer that ensures regulated data remains compliant throughout the analytics pipeline.

What are the three pillars of the Qlik and Databricks joint architecture for healthcare?

This video organizes the joint solution around three pillars: governed management of regulated data, AI-powered automation of diagnostic and analytical processes, and user-facing visualization delivered through Qlik. These three pillars correspond to specific capabilities demonstrated in the video, including near-real-time claims processing and AI-assisted equipment diagnostics.

Full transcript

[00:09] Good morning. Welcome to the session how data bricks and click can help health insurance organizations to makes their life easier. Um I would like to introduce my myself. I'm Christopher Christoph Murder. I'm based
[00:21] out of Germany. I'm working quite close with data bricks with Matias. um together I'm leading the data bricks business for for click globally and I'm
[00:33] also a specialist in data integration as well. >> Yeah, thank you very much Kristoff. So my name is Matias Matias Ingerfeld. I'm normally based if I'm not in San Francisco attending conferences in the beautiful city of Cologne in Germany. I
[00:47] lead for data bricks in central EMIA the field engineering organization. So it's a very technical organization, a lot of solution architects, specialists and delivery solution architects and um internally I'm called the tech GM so
[01:00] it's easier to remember versus long titles. Um and I'm happy to be here. It's a great session because we want to also show you a bit how the two companies work together in a practical terms and inspire you a bit uh what you
[01:12] could do things that we can solve jointly. Um yeah and I think it's it's a great event so far. been really excited so far and uh today last day full of action. So over to Christoph.
[01:25] >> Yeah. So what we would like to talk about today is how click and data bricks works. What what is the alliances? What is the ecosystem um together from data bricks also from click side and also the the use cases for health insurance
[01:38] organization. What is the benefit for them? Yeah, with that uh first um this is an old slide, but it's still true and very valid because uh Click did it again. So, congratulations Kristoff um
[01:52] and Click team. So, you've been recently winning awards in the past and fresh hot of the press. You have been also awarded this week on I think it was Tuesday. So, >> Monday was Monday. It was the integration partner of the year for for
[02:05] data bricks. >> So, congratulations. Maybe a round of applause. Thank you. Great. Um, let's talk more about what we do jointly together. And obviously, um,
[02:18] we we're long-term, um, joint collaboration and joint partnership and a lot of positive outcomes. And the two portfolios, I know both companies are expanding a lot what they do. You heard a lot of announcements this week, but
[02:30] they really match well together and and it's like quite a nice fit and many many use case. We'll talk about specific use cases later. Yeah, >> that's So first let me do it very short and as
[02:42] it's a data bricks conference so I I assume you've seen all this but I just want to give you a bit of a context um to why we're thinking this is really a good match and so obviously data bricks has evolved and you heard the data
[02:55] bricks evolution into the intelligence platform and now even moving beyond and you know our history I think your our background we've been doing AI for a long time I'm quite actually proud about it we were called the data and I company even back before 2019 and for the hype
[03:09] and this conference actually was also called now data and I summit for quite a long time previously it was then back then many years ago the spark summit but basically that's where the company roots
[03:21] are from and I think it's important to know that that the DNA of data bricks is essentially also heavily on the eye side and that was even prehype cycle so to speak um pre before the the times where
[03:33] all the LLMs and all the agentic magic came up so that's I think an important background Um because I think it's a bit like a dual track on a motorway. You have to do everything as a customer. Unfortunately, life's complicated in it.
[03:46] So you probably need to run what I call the truck on the on the motorway on the run one side where you have the heavy workload, you have data warehousing, you have a lot of uh very very important things on the heavy lift on the data side and also you you want to drive the
[03:59] sports car at the same time I guess. So you have also the AI part and the more modern way of doing things and ultimately you need to combine both and and it's many ways to interlock with the two companies. I keep it short um you
[04:12] probably know all this that Unity catalog um is the flagship product by data bricks to really give you endtoend governance and lineage and many many other things. It's the overarching theme. You heard it also recently in the
[04:26] keynotes that we're now extending this far far beyond as you're probably aware. We can also govern LLMs either internally on the platform but also external models. So it's it's expanding quite a lot now and you saw probably in
[04:39] Mate's keynote about cost control and those kind of things. So it's it's quite a massive portfolio by now. Um yeah and then jointly with click we're actually driving to this joint AI success and and
[04:52] I here three pillars that I listed. So obviously I'll start with the bottom. So the complexity of the AI integration as I said um it's not easy and usually you find in the way to do this you need at
[05:05] some stage get a really good grip on the data and as you probably all aware good data makes good AI. So, and often the companies have the good data not only in a nice shiny uh warehouse in the latest
[05:17] uh lake lakehouse kind of style. Often it's buried under many things in many different clouds, multicloud, crosscloud or on premise as well. And usually what happens then we need help from ISVS and
[05:30] partners because the real world is quite complex. So to do that end to end you need a strong partner that can do a lot of integrations on the way and also much more. more I mean you'll hear a lot about it from Kristoff. Yeah. Want to
[05:43] chime in on this some >> later on later. Um then also I mean we talk about healthcare life sciences. Uh the second point is close to my heart. For example, we published a blog from a German customer this this week or last
[05:55] week technically speaking about uh lake base and uh it's a healthcare sorry it's a insurance company. It's Argo insurance and it's not easy in a highly regulated insurance or in this case healthcare environment to really ace all these tick
[06:09] boxes that you need to navigate around every day literally and we're quite proud that we have a lot of regulated customers and I think if you can crack the code in a highly regulated environment then you can do it easily
[06:22] for dynamic startup that should be much easier to do as well. So I think um it's important to respect that uh and by the way we're also working in data brick side more and more towards achieving a lot of the industry standard certifications. It's a long journey.
[06:35] It's multi-year journey. Um but that's full progress now. We have a lot of things if you think financial regulation dora and other things in many industries we we're able to do it now. Um and the same for healthcare life sciences
[06:47] there's a lot of regulatory um items we are progressing on and we can fulfill them already. It's not like a a future story. This is a real story for today. And the first item uh I mean this is the
[06:59] story hunting the whole IT industry for many many years now. the silos are real unfortunately and um you need really good technology and I think it's often needed to think beyond the platform. I often introduce the le the data bricks
[07:12] platform bit like Lego blocks you have building blocks you can use certain things for certain things you can go end to end all in with data bricks you can also augment it massively with our partner ecosystem and our technology partners so and I think the customers
[07:26] that are really clever they usually pick and choose what to do where and actually they avoid a lot of overhead in time and consulting companies extra work and extra time um so my my advice is be
[07:39] clever check about what's out there on the market because you'll be surprised if you compare both portfolios next to each other. There's really a lot of magic integration and it's it's really matching quite well and it saves you a lot of time by the way. So I think um
[07:52] the the theme is don't don't overdo it. Don't rethink any sol problem that's solved already. Okay, cool. Um two more things >> and we have a good slide for that on later on that shows here where click
[08:04] fits in into the data brick >> yes >> environment. So I mean you're in a data bricks conference. I'll keep this short. Uh we're out there now around 13 years and been evolving quite dramatically. Um I think it's proven and obviously since
[08:17] the growth is real with many customers that we can really do quite heavy workload on on um really massive thousands of users we we often look at how many users have customers on a
[08:29] weekly basis on a monthly basis. It's quite impressive. um some have hundreds some have thousands but we have seen also tous thousands of of users on the platform and that's that's I think the magic I won't go into detail here um
[08:41] because you have probably so many sessions on a lot of things now uh obviously one thing to note and that's important for our partnership um as you know we did uh change the the relation that we have Iceberg and Delta so we we
[08:54] want to make this kind of industry battle with the two formats go away you probably heard it also on stage on the keynotes on Tuesday Um that's the goal. The company wants actually to merge these formats ultimately and in a good
[09:06] way to make the format a non-issue. Probably you saw that when one of the founding members of the iceberg format came on stage uh on Tuesday saying actually the the format discussion shouldn't be a discussion. And I see customers are less and less now
[09:19] interested in the format war so to speak. They're more into okay it doesn't really matter. We can do it. But it's interesting because a lot of customers prefer iceberg and it's yeah it's doable now. But say if it comes to cost
[09:31] question or seeing where your costs are coming from definitely we can feed as well also iceberg that's I think out of the box solution inside inside click and definitely where we can help you as well. Cool. Um and yeah then I have two more slides again I'll I'll skip them a
[09:44] bit in the awareness of time because we want to talk about the healthcare life sciences part but I just wanted to reassure that the both formats are very real now. So Delta has been there for many years but even more so on on the
[09:57] iceberg side that's highlight from my side that it's not just the connectivity that is is possible it's also performant connectivity and that's a hard work if you just do a connector it's okay but you want to connect really on high
[10:09] performance and of obviously it was massive transformation our portfolio that we had to change a lot to make our platform end to end iceberg enabled and also being able to write not only read that's often the magic and where the
[10:21] complexity comes comes from and I think Click has many years of experience of on that on that theme. So, uh that's that's a theme you know quite well. I guess Kristoff, >> that's correct. >> Okay, cool. Um yeah, final thing not as
[10:34] a reminder um unity is everywhere and we can plug that nicely in with for example click and other technologies where in the end you want highly governed data and um it doesn't matter also how you do the visualization layer. You can use for
[10:47] example elements from click there. You can also use elements from click on a connectivity side. We see where it's really on the practical level very powerful. I give you two examples. You can connect to mainframe with click which we cannot do as data bricks
[11:00] without the help of an ISV. We're not the software doesn't run on premise as you're probably fully aware. It's only running in the cloud in um either AWS, Azure or GCP. And um that also means
[11:13] essentially the more it goes on uh complex connectivity on premise SAP related topics it's it's we do need help and the help is next to me here right >> cool
[11:25] great um yeah Delta live tables I'll skip that I I think we have not this is a very powerful summary and normally when I present I have a slide with unbelievably amount of icons here uh I I
[11:38] I was hiding them here for this presentation but The special thing about click is um that click can cover a lot of these segments. So you see here from business intelligence, data governance and security, data integration obviously
[11:51] and a lot of other things. I didn't list everything here, but typically if you if you want to tap into the ecosystem, you end up with a lot of different partners. And this is also where I think um we can saves you save your nerves, time and
[12:04] your own like complexity reduction a bit in a in a way that you do not pick and choose for every single task either our own platform in data bricks or uh an additional um as we say technology partner ISV. In this slide here, you see
[12:18] that actually click can cover a lot end to end and that's super unique. Normally we have partners. They're very good at one singular pillar as I call it and then when it goes broader it's usually
[12:30] yeah pick another one, pick one more, pick one more and you end up with four, five, six different technologies if you have to expand into a complex scenario. So >> yeah. >> Yep. >> That is correct. And usually let's say where click is coming from is on the
[12:43] business intelligence side. So click found a couple um a couple of decades before so 30 35 years something like that and just started with click view then with clicksense we have clicksense cloud that's a product for visualization
[12:56] of data through the years there are quite also some companies let's say for data integration that this piece here they um acquired a company called a real-time change data capture pro um um
[13:10] um uh product and what you said before is If you want to connect to mainframe or SAP definitely we can do real time change data capture into data bricks and that's quite important depends on the use case that's that's also also really
[13:23] important um you have use cases which is not really relevant to have real-time data but in lots of um projects I'm running definitely it's a big request to have such data let's say in a couple of seconds in in in um in data bricks
[13:39] through the next acquisition we also enhanced our product where we also acquired a company called talent so where we also integrated the data governance data lineage data trust
[13:51] interet whole product so what we are doing with click we taking your row data and make your data AI and analytics ready that's a big process in in between it's a pipeline process first it's an
[14:03] ELT the loading part and then we switch over automatically in the ETL part to transform the data in the right way so it can be used by data breaks.
[14:19] A few words to the click talent cloud platform. What is um important on that slide at all? We make your data AI ready. What does it means and also what I said before is bringing data into the data
[14:33] bricks platform that is quite important. We call it the data foundation. We need to have qualified data which can be used by data bricks and that is definitely what we are doing with a change data catchup and product inside click on that
[14:47] whole process also we we enhance the data with quality trust linage all the transformation we can add to the process to make a good qualify out qualified
[14:59] outcome for your data and then of course we have click started with BI we have the analytics on top of it we have also our own let's say AI product uh integrated in in the click platform you
[15:12] can use let's say in a project we also can combine with genie whatever you you prefer to do we can put on an MCP server on top of it to get the best outcome for your queries you're sending um to data
[15:26] bricks as you can see here we have >> yes a lot of joint names >> yeah joint names and lots of really nice wins we we together. Last year we did
[15:38] also a presentation talking about Mercedes-Benz. It's also a really nice use case where we are a good fit for for for that company >> and you see the list here. It's quite impressive I would say. >> Yeah. And those of you visiting our
[15:50] conferences so there's a lot of overlap and the people also presenting and and showcasing on our conference or joint conference sometimes or also on the click conference like click connect. So there's there's a is a powerful um
[16:03] portfolio and also challenge us and come to us after the show if you like if you want references talk to another customer. I think it's always powerful to talk to someone who's done it or solved the problem. I think if you have
[16:16] lots of data in different systems and you would like to move it into into the cloud and specifically into data bricks and definitely the fits together is perfect I would say because we can help them to get the data in transform the
[16:28] data in the right way and make it usable for a wider group inside an organization. Just a few few a few numbers about what what what we having behind the product
[16:41] and change data capture is one of the most important stuff as you can see here we have 400 plus different connectors it's not really true so we have 100 connectors behind the firewall that
[16:53] means connecting to SAP mainframe SQL Oracle whatever you have definitely we can access the data and bring it into the databicks platform and also our let's say multi pattern we can let's say
[17:07] we we have two different kinds of let's say um projects inside the pipeline pro process first we start with the ELT the loading part and then we switch over automatically into the ETL part
[17:21] on top of it we have lots of connectors for cloud applications like service now workday whatever you have inside your organization definitely we can add the data so we have an API call to also extract the data quite easily
[17:34] Our transforming part it's we are saying all the time no code low code we're trying to make it as easy as possible. If you would start learning or working with click you don't need to go to to a
[17:46] classroom for two three weeks definitely after a few days you are able to to transform your data in the right way. Quality is also a big part of it. When you access data you see we can measure
[17:58] some trust score. Is the data really valid for you to move it into into data bricks. We can have kind of a score ranking. Is it a good score? If not, that we can adjust the data to make it more usable later on inside data bricks.
[18:12] Our key um play together if you have any SAP and mainframe projects definitely definitely it's a key play for for both companies and it's quite important. Yeah, I might to highlight here on the SAP side as you're probably all aware
[18:25] from your own companies that SAP is quite large and complex and even if you have parts of it on the latest and greatest edition essentially often in a in a rebuilding renovation and renewal phase where a lot of the platforms have
[18:38] to be made fit for the future. So what we see with customers is that even if they have one part in the absolute top state which is um latest greatest um fully on the road map of SAP there's
[18:50] often a lot of a mix under the hood where you you can't do everything at the same time. There's huge transformation projects going on for a long time and the magic here with click together is we can address all of that typically which
[19:03] is quite unique because normally there is a connectivity to one addition one one thing and you know we've been launching a lot of news on that last year. Um but essentially most of the environments are quite distributed and
[19:15] complex and and there not only nice and shiny new there's also very old and then how to tackle it without asking 15 companies to help you individually that's that's where you can go to one place and can cover it with just us and
[19:28] click together >> also we do support AI our let's say internal AI we also can connect easily to to to data bricks or the other AI vendors in the market and also putting MCP on top of it to query let's say all
[19:41] different kind of sources you would like to to do. ISPEC support we also already talked about that and we have also four different kinds of additions of of the click talent cloud platform. Just to finish that quite quickly um
[19:56] click is now as I said before is more than 30 years on the market 35 years something like that and definitely we know what we what we are doing starting let's say with a change data capture product um we can easily connect to any
[20:08] source on any target and moves the data in into data bricks that's I would say it's yeah that's our standard I would say next to the movement we do also just a transformation but it's quite important um we can integrate it um
[20:21] trusted um quality checks etc. What I said before and also the good thing is we can when we do the replication we can do structured unstructured and semiructured data to move it directly into the cloud.
[20:34] um agile we we are I say cloud vendor independent so we are running let's say similar to you on on AWS Microsoft and and Google um of course we are also
[20:47] working with snowflake and and the others it's definitely we are also independent in that area and also we can integrate AI from every other vendors as well and enterprise grade um the CDC product
[21:02] came out on the market 1991 on starting with main mainframe replication and and some other ones vom whatever is available definitely we can uh we can do
[21:14] CDC and we we are doing it for many many years and definitely we are enterprise ready and we are also scor on many on gardener on on other on the top magic red >> I would like to highlight the MCP story
[21:27] so you know that the industry is moving obviously everything is MCP ready now you can connect everything and it's for both companies very similar approach you heard it probably on the keynotes that Genie and MCP is is a good fit but same
[21:39] here and that gives you a lot of options how you want to set it up so it's it's quite powerful that the two companies use the same standards and you can build it on data bricks you can use also elements with click and and because it's
[21:52] all like nowadays MC MCP connectivity so that's where both companies going in the same direction from product portfolio so that's very powerful actually we've seen customers uh do that already they they do pick and choose and I think choice is
[22:04] important. You heard also um we're we're very open and the choice is yours and that means also on the ultimate question on AI it's your choice which platform you take in the end >> and also from the MCP side we also
[22:17] listed on on your marketplace. >> Yes. just a quick I think one of the last slide talking about the product um just a quick overview how we let's say work with with data bricks together started
[22:31] again for me when when I'm going to a presentation to to a customer or a partner I would like to ask the question how you would like to start using your data what is your typical use case and
[22:44] and I need to understand what you would like to build and then I go to to the left side and say okay I know what what you're looking for can you tell me please something about what kind of source data you have what is the structure of the data what you would
[22:56] like to build with all the different structures should it be done with inside data bricks should it be done in with with click do we should do a kind of a combination that's definitely I need I need to ask lots of question to get the
[23:08] right let's say um yeah the usage what what you would like to do and then we start the process and start with real time um change data capture so We move
[23:20] the data into a landing area inside datab bricks. It could be let's say on on S3 for example, could be also in in data bricks directly and then we start let's say you're doing also transformation.
[23:35] Yeah. and and my best example is all the time you have a customer ID let's say in mainframe you have a custom ID in SAP maybe another c customer ID let's in a SQL database we can start defining a new
[23:48] data structure new data model which what we are doing then we do a kind of a push down SQL into data bricks and then we start storing the data in the right format which can be used um later on
[24:00] with Genie and and and as well also with click applications On that process we also do the enrichment of the data with quality trust governance lineage etc to provide
[24:15] you the best outcome of your data. >> Maybe just to add also I spoke to a number of customers this week and most of them say when they do transformational projects they they don't want to obviously lift and shift from A to B. They want to really
[24:27] transform the data and have a really high quality data and day zero when they start with data bricks. There's lots of things in the platform with click to improve the quality but I I think a lot of them do it in a clever way where they
[24:40] don't just move what they have. They really take a clever look okay what what's the high standard what do we need what not need and then uh actually it can help a lot at the initial transformation when the data comes in and I've seen that's much easier saves
[24:52] you a lot of headache later because then you don't need to uh clean that up later. Yeah. >> Yeah. And also between the two products there's also sometimes an overlap and we need to find out what is the best way to provide the best outcome for our customer. All
[25:06] right and then that is different kind of use cases um just pure dashboarding the process looks almost the same I would say we have machine learning we have also said generative AI the the data
[25:20] foundation is in all three the same and the outcome what you are doing let's say here that's a little bit different and also how to provide the data in in in a in a dashboard in an application that's definitely something needs to be
[25:33] discussed um for every kind of use case. >> Maybe also a practical example where um you can save time um doing things today already um as you know probably leg flow is evolving. There's lots of more and
[25:46] more sources coming into a data bricks part where we can uh ingest already directly from lots of sources. What people often don't know and I'm very sorry for them because they spend a lot of time searching they don't need to search uh that is out there already for
[25:59] many years with click. So my example is you can connect to SharePoint I think for more than 10 15 20 years or or many years and uh we we came out quite late on this and so sometimes it's probably the better way to say instead of waiting
[26:11] for a feature to come up just look what's there likelihood is very high it's it's already covered and then it's covered in all flavors in all editions and so on because most companies when I
[26:23] say for example SharePoint or other things they they have all additions they literally anything and everything from on premise to cloud and and then that makes it complex and uh that is usually I see that on the architecture diagrams ending up with many errors so we can
[26:36] save you a lot of these errors >> I think connecting to SharePoint maybe is a good example um but I think a different version of course I know that but also loading data is one piece the other thing is what are you doing with
[26:48] change data and that is definitely something let's say take as an example we connecting directly to what's called transactional lock to see if change is coming in if change is coming in we takes the data move it autom automatically in real time in into data
[27:01] bricks and that is the challenging part and we doing for 100 plus different sources in in the behind the firewall and that I would say it's a a big plus of click I would say now to the
[27:14] healthcare use case um I think it's let's say I will start and hand over to you and usually we have a problem in lots of organizations they have not only one SAP system they have
[27:26] maybe machine data other datas you need to get all the data and move it as I said before with all the transformation included into a modern data platform on
[27:38] data bricks and then you are coming into that that game and preparing the data. >> Yeah. Yeah. Maybe I can give some practical examples. So we've done a lot in in where healthcare data is literally
[27:50] physically being created and then obviously it's going to be part of the electronic healthcare record later. So let's say for example you do an MRT or a similar things and it produce a lot of um data that nowadays all is electronic
[28:03] obviously and then you know um it's it's a challenge because it says here on the slide near real term real term let's say you're at the doctor and you you do a certain medical examination and then you go to the next doctor next door and the
[28:15] data is supposed to be in there already and now nowadays it's it's all almost near real time or real time because there's no time to wait you can't wait for documents to be sent anymore I mean this is the old way of doing things. The industry has transformed a lot and that
[28:28] that brings also a lot of volume complexity. So imagine your healthcare vendor bringing the machines in that do some patient diagnostics. So that could be absolutely massive in terms of
[28:40] volume. We hear later in a specific use case a bit how high the volumes are, but we're talking a lot because nowadays because you can capture and grab all that information. It's there. It's possible. Why not capture as much as you
[28:53] can at source and obviously then we're in a highly regulated environment. So um then comes the second section here into play that the delivery needs to be governed end to end controlled and also highly regulated. Often you can't just
[29:06] store patient data in a in a wild west style. So that's uh where typically has two dimensions this problem. One is at source where the source produces an overwhelming amount of data. Think about
[29:18] images even videos now and lots more data is being generated in in seconds and it has to be digested so to speak and then secondly on the way end to end it it really needs to highly have a governance model but it's complex
[29:31] nowadays you could have even the the provider of a certain um diagnostic machine could actually give you the service. So it's technically speaking a third party. Then you digest it. It's a clinic let's say. Then the clinic is
[29:44] operating operating on the data. They're they're actually responsible for the data in that case because they're they're using the data and um and then it could go to your patient record and something like that and could be to going to another clinic to another
[29:56] doctor. So the the chain gets quite complicated and nowadays a lot of the data is also um provided in a service manner meaning um you have like input outputs. It's sometimes also with AI
[30:08] that um some of the vendors do part of this stuff. They would analyze the pictures for you, send you back a result or pre pre-assessment and stuff like that. So, how do you actually combine that together? So, um if I would draw it
[30:22] now, it's it's a picture could start in that end and end in the other corner of the room and and in between uh if something breaks is problematic. It's dramatic. So you need to be really aware of the regulations and that's why the
[30:34] platform is end to end governed and not only partially governed. That's the main plus. Yeah. >> Yeah. >> And usually how it looks like let's say as as I talked before and have in a
[30:46] different style um you have different kinds of systems. Yeah. As you can see have a hospital system you have some um machine systems etc. You have also some other medical devices also from inside
[30:58] or outside. we can also access all the different let's say data and move it I call right now to a central platform it could be inside and outside of of a clinical um so that is definitely how
[31:10] you would like to set it up and then move the data in and also bring more value to the data do it with with with with governance rules and etc what we can add to them to in the
[31:22] movement process to get a good qualified let's say outcome let's say for for data bricks >> one more addition to This one if you look on the right hand side how data is consumed has dramatically changed in data bricks in the last two three years.
[31:34] So previously if I was here like 5 years ago uh probably it would have been the main question how can technical people actually digest the data you know when you go in the data platform usually you see a very technical guey obviously you
[31:48] see code you can work with code data scientists data engineering ML people can work with it and what has really dramatically changed is now also on the bottom right hand side that you want often direct access from people that are
[32:00] more either executive business people or they're not programming experts they they're not technical audience that that use the platform and you might know we have also data bricks one we have Genie and this is happening there on the right
[32:12] where a lot of the visualization and the work with the data is talk to your data nowadays it's not anymore program code with your data so to speak and and this is uh really completely changed our our user base as well so if I look at
[32:25] statistics we work with our clients we look at how how is your user base structured what is your typical end user typically it was a technical user that is not scared about code seeing a few lines of uh code or SQL queries and so
[32:39] it's no problem but now it has dramatically changed as people their business so in the healthcare context they they want to know things high level they want to know for their business they don't know really how it works behind the scenes and and that's where
[32:52] the visualization in the end is is either in data bricks one which is an an for more less less technical more business audience uh or also in Genie where you can simply talk to your data you ask in normal plain English even in
[33:05] any other language your questions and you get answers. So, so that's where the dramatic shift happen and this only works if it's end to end and the challenge is before with with some vendors this challenge is that uh the
[33:18] the chain is broken in a way there's parts where some vendors are really good but end to end is usually so so complex that it's a tough one to crack. Yeah. And also you must be on that side here and that's my question when I go to to
[33:30] our client or partner and saying what is your your typical use case and you must be also flexible to add another five 10 maybe 20 use cases um connecting to the platform and also getting the best outcome um for you. Yeah,
[33:43] >> that brings me to the next slide and it say something um company calls or seammen's health or it's called in earlier time seammen's medical systems. Um so let's let's say what what we are
[33:55] doing we are moving um 100 terabytes plus in into the data bricks platform um um machine data um documentation and etc. All the different
[34:07] kinds of information are stored let's say in the in in in the data bricks platform and usually on the analytic side and this is quite impressive. So if the management would like to have a better control of of what's going on in
[34:20] my hospital then we have also let's say the seams would like to see can we add more services to that to that hospital as well and the third big use case is for for the engineers and usually if something happening with machines and
[34:32] MIT system or CT system you're running into an error and usually with that platform that's quite genius I would say because we included pictures and videos um the the engineers the typing in the
[34:46] error code and the first what is getting how to fix that machine step guidance um to makes the machine working again and it's a big time reduction um um or more
[34:58] engineers more efficient than than before and that is quite powerful I would say >> I would like to add the cost element that was also on the previous slide here what also often is underestimated so the the licensing model in data bricks
[35:10] you're probably aware is is mainly compute usage when you actively use the data you pay. And so what customers really love is they they kind of pump the data and you see it's massive amounts. It's it's a really this is just
[35:24] one example. Imagine you have 50 like these use cases. You have other use cases in in the same company. So the the volumes are are really going up dramatically. But what customers often do now, they're very selective in the end what they actually do with the data,
[35:37] but they want to capture it. So they don't want to filter at source where because of the the cost that it produces later that uh oh I have the data but actually I need to drop half the data because I'm concerned about the cost. That's actually a weird strategy to be
[35:51] honest. So the strategy is nowadays uh use everything you have and even data points that you might think are not so relevant, you just still capture them. um you might need them later and and then it's actually if you don't use them
[36:04] if they're just sitting let's say in the in the storage bucket in the S3 or similar stoages it's relatively cost neutral relatively cheap but then it's being worked on then actually um you can do very powerful things and you get
[36:17] value out of it so so the cost value relation is quite good and often we also see customers eliminating the complexity of their portfolio they might have it today already but it's a it's an endless chain and there's delays there's risk
[36:30] there's complexity and they kind of shortcut the whole chain end to end with the two companies. >> Yeah. >> And on the visualization side, I think you you talked quite a lot about Genie. Of course, we have also click has also
[36:43] click sense. So definitely there's no it sounds like a competitor or competitive product maybe yes maybe no but in the end we also can work together with with Genie and click apps. It's
[36:55] also I think that's that is a clear message. I would say if you have some specific let's say usage for for data definitely um clicksense is a a perfect product. Also we can go combine it also
[37:07] with with Genie with with MCP what what we talked about before. I think that is also the message I would like to deliver. >> Yeah. And it gives you a choice. I mean that's uh it's up to you what you want to use. Some customers also have a lot
[37:20] of experience for many years with click. They prefer the click side. Some are ramping up on the data bricks knowledge. So it's not a either or this such discussion is very important. So you easily tap into the the both portfolios. I mean why would you not do that? It's I
[37:34] think the more clever choice. Yeah. >> Good. >> I think we are almost >> with that. Yeah. We're we're done with our part and we're here quite a bit still. So feel free to come to us. We have any questions. Um we're very
[37:46] thankful that you spend the time with us and >> should we refer to him?
[37:58] being here. Was a pleasure hosting you and it was fun. Thank you so much.

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.