Skip to main content

Customer 360 with LakeFusion MDM and Dun & Bradstreet Enrichment on Databricks

Summary

  • Ebara Elliott Energy reduced customer records by 50% through AI-driven probabilistic matching using LakeFusion MDM's Match Maven engine, native to the Databricks Data and AI platform, resolving duplicates across multiple ERP, CRM, and billing systems.
  • Dun & Bradstreet enrichment added corporate hierarchy, segmentation, and revenue intelligence on top of deduplicated records, enabling use cases beyond sales including compliance, credit decisions, and AI-readiness.
  • Human stewardship workflows were essential to production-quality master data, as automated matching alone is insufficient without domain expert review and a mindset shift from siloed data ownership to collaborative data governance.

Customer 360 with LakeFusion MDM and Dun & Bradstreet Enrichment on Databricks

Watch: Customer 360 with LakeFusion MDM and Dun & Bradstreet Enrichment on Databricks
Manufacturing enterprises face fragmented customer data across multiple legacy ERP, CRM, and billing systems, creating duplicates, inconsistent reporting, and missed sales opportunities. Elliott Energy struggled with incomplete customer pictures, duplicates within and across systems, missing hierarchy information, and poor data quality. By implementing LakeFusion native master data management on Databricks combined with Dun & Bradstreet enrichment, Elliott reduced customer records by 50% through AI-driven matching while preserving data quality through human stewardship.
this video demonstrates how to build customer 360 using LakeFusion MDM's Match Maven AI engine for probabilistic matching, combined with Dun & Bradstreet's commercial graph for hierarchy and segmentation. You'll learn governance architecture with stewardship workflows, how Unity Catalog enables lineage, and adoption strategies for cross-functional alignment. The session covers use cases beyond sales including compliance and credit decisions, plus how to prepare data for AI applications.
🤝

Chapters

FAQs

What is LakeFusion MDM and how does it work on Databricks?

LakeFusion is a native master data management solution built on the Databricks Data and AI platform that provides AI-driven matching, data stewardship workflows, and governance through Unity Catalog lineage. Its Match Maven engine uses probabilistic matching to identify and resolve duplicate records across multiple source systems without requiring exact field matches.

How did Elliott Energy achieve 50% customer deduplication?

Elliott Energy used LakeFusion MDM's Match Maven AI engine to identify and merge duplicate customer records that existed within and across their ERP, CRM, and billing systems. The 50% reduction in customer records was achieved through a combination of automated probabilistic matching and human stewardship review to validate merge decisions.

What does Dun & Bradstreet enrichment add to a customer master?

Dun & Bradstreet enrichment adds corporate hierarchy data, firmographic segmentation, and revenue intelligence to deduplicated customer records using the Dun & Bradstreet commercial graph. This enables use cases such as identifying a customer's parent company relationship, targeting by segment, and supporting compliance and credit decision workflows.

Why is data stewardship critical for master data management?

Automated matching produces candidate merge decisions, but domain experts must review and approve these decisions to maintain data quality, particularly for high-value customer records with complex relationship structures. This video emphasizes that a successful MDM program requires a mindset shift from siloed data ownership to collaborative stewardship across business functions.

Full transcript

[00:08] Awesome. Well, end of the day, welcome to um a session. This is a three person session, which is rare for Databricks. Um we're going to introduce ourselves here in a minute, and then what we're going to be talking about today is Elliott Barrett Elliott a Barrett Energy's journey through customer 360 data with Lake Fusion,
[00:25] which is a Databricks native master data management solution, and Dun & Bradstreet, which is um data enrichment investment class data enrichment. So, um Madhu, will you introduce yourself, and we'll listen to myself. Hi. Hello, everyone. Good evening. Um
[00:40] it's amazing that it's been a long day, and it's 5:20, and you're all still here. That means you're very, very serious about master data management. So, um can I get the clicker? Go for Okay.
[00:59] So, you want to talk a little bit about this, or we move on? So, the purpose of what we're doing today is um talking through the Elliott a Barrett journey. So, um Madhu here um is experienced and she'll go into her introduction in a minute, but we'll talk through kind of what their story was with master data and with enriched data. We'll talk about
[01:15] how Lake Fusion and Dun & Bradstreet helped, and we'll wrap it up about what's next. And I think that's going to be really helpful for everyone cuz Madhu will go into some of the challenges that she's faced, and then also some of the things she's dreaming about doing in the future. Okay. So, my name is Madhu Madhu Kudaravalli, and
[01:33] um I've been in this industry forever. Uh I'm I'm really a data person. I'm passionate about all things data. Um been working with data for all of my career, I suppose. Um
[01:48] this and I I was this morning I was trying to think about, like, you know, about this presentation, and I realized this is my fifth master data management project in all of my career. Um, I've built two in-house,
[02:07] um, worked with very large companies which I don't think I should take all the names, but these are really huge big companies that were very very popular. Um, there was not much competition in those days with these big companies. And
[02:23] now the fifth with Latifusion. Um, I highlight the banking and health care over here because these are industries where know your customer is so big, right? People are so huge on I really want to understand the green
[02:40] space, some call it the blue space, whatever the blue sky, but that's what it is all about, know your customer. For us as a manufacturing industry for which I work right now, it is more than just about the customer.
[02:56] And we'll talk about it a little bit. Um, so I work for a company called Ebara Elliott Energy. And what is Ebara Elliott Energy? Many people call us a manufacturing firm. Yes, we are a manufacturing firm, but I
[03:12] consider us more of a design firm. Because what we do is something very niche in the market. We don't have too many competitors. And um, we are known for the quality of the turbines and the compressors and the custom pumps
[03:30] that we manufacture. So these are not like, you know, the the regular run-of-the-mill. They're made-to-order. Takes, you know, months and months for design and then manufacturing. Like sometimes I've seen trucks like as long as this
[03:45] hall. That's how big the turbine is. So we're doing some pretty cool stuff. But we're also very global. our parent company is in Japan and the parent company is called Ebara and it has five verticals. And one of
[04:01] them is the energy vertical and that's what where I work for. And we are even though we are headquartered in Japan, one of the manufacturing So, all the red dots you see on the screen, those are all our manufacturing units. And
[04:19] um I work at Janit, which is in Pittsburgh, Pennsylvania. Uh go Steelers. Um if any Anybody here? Oh, great. Cool. So, um So, I'm from Pittsburgh. Um and uh this manufacturing firm is
[04:37] almost, I've heard, 125 years old. Uh earlier it was just called Elliott Turbo and now it's part of the Ebara group of companies. Um So, as you see, we are global. So, our customers also are global, right? So,
[04:55] it's very important that we focus on our customers and have a true picture of our customer. So, I'm going to start with a story of the elephant and five blind men. So, some of you may have heard the story,
[05:11] but you blind five five men have, you know, blindfold. One of them touches the trunk, thinks it's a python or a snake. The other one touches the tail and thinks it's a rope. Someone touches the fan, thinks
[05:27] uh the the ears and thinks it's a fan, right? So, you know, someone touches the body, thinks it is just a wall. So, everyone has a different perception of the elephant, right? But, no one
[05:43] knows the true picture. What animal is this? Right? That's equivalent to our customer. No pun intended. Not that the elephant is the customer, but if you take the analogy, you really don't understand your customer. So, we have partial information across
[06:00] all of our systems about our customer. Very common. I've I've seen this for years and years in my career. The partial information provides us this assumption that this customer is this small, they only work
[06:17] in this segment, you know, this is all they do, and we miss the board on how big they are or what industries they're working on so that we can tap into the green space, right? So, we need to know
[06:33] the truth about the customer, and that's where this master data management becomes very, very important, right? So, you know, now this is the practical, you know, issue we have. So, we at Elliott,
[06:49] like I said, we are a very, very old company. We have tons and tons of ERP systems, like a lot. Um I don't even want to tell you how many because it's it's laughable to a point, but this is because we have
[07:06] acquired and merged with companies, and so, you know, this is a common thing, but we can't solve all of these problems, right? So, we have to find a way to solve the truth, right? To find the truth. So,
[07:22] there are duplicates across the same system, across multiple systems, and there's no clear industry segmentation, which we want to know about the customer. We um You know, a very common problem that
[07:37] I've seen, and which is what where Dun & Bradstreet comes, is the hierarchy of the customer. So, for example, Shell we work with Shell. So, the way right now our folks, our sales folks will look for the word Shell and assume that all of
[07:54] these fall under the Shell umbrella, right? What if something merged, but it is under the Shell umbrella and doesn't have the word Shell, right? So, you really do not know the parent and the child relationship of your customer. You
[08:09] have them individually, and sometimes you're reporting sales saying this is a separate customer, this is a and I've seen this all across, like we do that right now. But, if we want to say like what how much revenue are we making from Shell, we need to know, hey, how do we
[08:25] combine all of this data, right? So, the other thing that people lose when um, you know, uh, not think about is in some systems you have really good information related to, let's say, billing. But, in
[08:42] another system you have all the corporate information, which is really good. So, you create a single record, which gives you the breadth of that customer across all of these systems. So, this is what we are trying to solve.
[08:58] But, we missed the bone on saying that, oh, it's mostly for sales. We care about the customers so that we can make more money. But, in our organization, we use it for so many different things. Not just sales, compliance. Now, with
[09:16] all of these, you know, regulations, sanctions, we are not supposed to work with certain customers that we used to work with. But, every year that list comes up and we have to make sure we're not working with them anymore, right? So, that's the
[09:33] compliance. And And contracting. We end up having contracts with multiple customers thinking they're multiple, but we apparently have a MSA with them already just because we have a different name. Right? So, these are
[09:50] additional areas. Otherwise, it's just know your customer, but know your customer for what purpose, right? Um marketing and corporate development, this we all know and even like after sales support, this is a common you know, um use case of this.
[10:06] And of course, the dashboarding reporting that that is a obvious and a given. So, why are we here on the stage, right? Because we use Lake Fusion and Data and
[10:22] Dun & Bradstreet. But, this architecture is to really represent how we started our journey, right? We wanted to keep our architecture very simple. Everything revolves around Databricks,
[10:38] right? We don't have a EDW, uh enterprise data warehouse, or nothing. This is it. This is what we have. We have Databricks, the data lakehouse, and we bring in all of our source data into Databricks. We use it for dashboarding.
[10:55] We use it for all kinds of analytics. We use it now for AI because Databricks serves all of these AI models, and we pick and choose which one we want to use depending upon the use case. So, we are a Databricks shop. That's how I would
[11:10] like to call ourselves. So, when we were looking for a solution to master our customers, our products, our suppliers, our accounts, there's a lot to master. We first started with the customers because that would give us the biggest
[11:26] bang for our buck. We started looking at multiple vendors. And like I said, we needed a one-stop solution that would work for all of the domains, not just
[11:41] the customer domain. Right? And we just wanted to make sure that it works with Databricks. Lot of the other solutions, you had to move the data, you had to stage the data in a certain place, and then master it.
[11:59] Here, all of our ERP systems, our CRM systems, all of the customer data was already in Databricks. Right? So, we wanted a solution that would work for us on top of Databricks. So, and of course, cost is always there.
[12:16] Support, a good vendor is very important. Learned that all through my career. Support is very, very essential. Doesn't matter how shiny the tool is. Right? So, we picked Lake Fusion, and we have
[12:32] been working with Lake Fusion for I I think probably around August or September of last year. Um and the when we if you look at the statistics, and I cannot share the exact numbers, but
[12:50] we decreased our customer master from like not customer master, we had let's say 100,000 customers, we um we brought it down to 50%. So, we
[13:05] we were down to 50,000 customers. Now, no tool is going to give you that you know, that magic match. Right? So, we did have like around 5% that was above the 90% threshold. Sorry,
[13:22] below the 90, between 80 and 90% match. So, that 5 to 10% we gave it to our end user, so there's a human in the loop. The rest of it AI was matching Lake Fusion was matching automatically. The rest of it our end users, there's a UI
[13:39] and they would go and they would match all of that. So, this worked very well for us because um if you look at it,
[13:57] the UI is built on top of Databricks. All the notebooks and the um so, you think about it, we have all of these ERP systems and we have our Databricks data lakehouse. Then we have Lake Fusion MDM which is helping us master all of that, but now we have
[14:12] already started mastering sites or like, you know, because we have equipment all over the world. So, and we have to match our customer to the sites, so we're doing that also. We are actually mastering our sites. So, next we'll be working on other domains.
[14:28] Um but then we're using this master data for reporting, for dashboarding, for AI, sending it to other systems, um and so on. Um we'll talk a little bit about Dun &
[14:44] Bradstreet later. Um but now I think I'm going to hand it over to Lake Fusion for presenting their part. Thanks, Madhu. I appreciate it. So, we're going to go through what Lake Fusion is quickly, kind of how we've
[14:59] helped um and how we've worked with Oleada Barra Energy and then we'll hand it over to Liz for Dun & Bradstreet. So, Lake Fusion is a data product. We currently have three products. Our flagship product, our master data management product, as Madhu mentioned, is built natively in Databricks. That
[15:15] means that all of the processing is happening in Databricks. It's a multi-domain MDM. You can master um domains, master relationships between domains, which brings us to one of our newest products, which is our lake graph product, which we can take that information that we've just mastered,
[15:30] those domains we've mastered, we can take those relationships that we've mastered, we can project those into a graph database. It's built in lake base. We can start to run contextual queries against algorithms in that, and we can do that with natural language, and we can plug that into Genie. So, it means
[15:47] we can ask questions about who's my most impactful supplier. If this specific part has a failure, what are my impacted finished goods? Who are my impacted customers? And we can apply that to all the domains. We can use what we've built in MDM to seed that graph. And then, the
[16:03] last product we have is a PIM product. This is product information management. We can use this to manage product or assets, um give it a single pane of glass so business users can come in and update information about these different products. Um I do want to show our architecture again, just to kind of highlight that we
[16:18] are built in Databricks. So, when you're clicking a button or UI, you're performing an action with a workflow or a pipeline in Databricks. That means the data is accessible to you. If you want to do anything with that data, put it into workflows, put it into triggers, you're more than welcome to do that. Um those
[16:33] notebooks are exposed to you in your catalog. They're governed by your Unity Catalog permissions uh within the platform. So, um something else we're really proud of at Lake Fusion is we know that in master data management, one of the hardest things to do is to match tune. Um there's two really hard problems in master data
[16:50] management. One is figuring out what usable data means, and then two is getting to that definition of usable data with your match tuning. So, we have a match tuning process. We're proud of it, we call it Match Maven. Um it means we do rule-based, so that's a standard traditional MDM um field-to-field mastering or or uh
[17:07] probabilistic matching. Then we also have a contextual match. We use Databricks foundational LLMs to look at a full and complete record, and then review it the way a human would it, and say, "Are these things similar? Give me a score on how similar these things are. Um what that does is allows us to take
[17:22] more time away from humans and put more time on the compute. Let humans do the hard hard matches and let us handle um the matches that we can now address with AI. Um and lastly, we have a really iterative process for how to do this. So, we have a playground. We can play with different rules, different prompts,
[17:38] different records that that are in the system. We can create fake records to see how it handle things like in-person mastering twins. It's a really difficult problem to solve. We can generate twin type data and then run it through our match engine to see how it would work really quickly um iterate those before we expand it to our large
[17:53] volumes of records that we're going to then have to process for a longer amount amount of time to get the results from. And then, last thing I'll show is um as Madhu mentioned, we have a stewardship UI. We have a card-based UI. So, we can see golden record. We can see the inputs from various sources. We can
[18:09] see what fields are coming in from those sources. This makes it really easy for stewards to review, and it's um kind of gives you just like a nice visual clean view. And then, on the right, that's the Data Bricks lineage of that master data. So, um just again, want to hammer this
[18:24] point home that we are built in Data Bricks. And when you're looking at our master data in our beautiful UI, you click a button, you can open the lineage diagram in Data Bricks to see where that information came from, and where it's going to in your systems.
[18:43] I think we're next. We're going to Dun & Bradstreet. So, we'll talk a little bit about enriching data. I think this is this is extremely important because how much data can we store in our I mean, even though we have tons of systems, honestly, there's not enough for us to
[18:59] get the true picture of our customers, right? Um because really, no one cares while they're entering things into the system. The financial people, all they care about is the billing address. That's all they care about, right? And then
[19:15] the the sales people think about like the the corporate address or some contact that they want to know and a phone number. That's all they care about. But if you really want to be strategic about your company and really look at what are the possibilities that
[19:33] you can have with your customer? What is the potential with your customer? You need to know more than what you actually possess, right? And that's where Dun & Bradstreet comes into the picture. And I've used several companies. I mean, for
[19:48] health care we've used Definitive. But for all other like non-health care I've used Dun & Bradstreet even earlier in my other MDM projects. And one of the biggest things that we wanted to solve was the hierarchy of the customer, knowing the parent-child
[20:05] relationship of the customer, which fall which companies fall under this one ABC company umbrella, for example, right? So we are getting the hierarchy from D&B because it has you know, the global hierarchy, the
[20:22] parent hierarchy, the child like it has all levels that you can go through and you can pick and choose to what level you want to go to, right? Um And then um one of our biggest challenges was the segmentation
[20:38] or the areas that the customers work in. And so what happens is when a new project comes in or when a new sales order comes in, we just look at the type of project and just classify that customer to a certain segment. Turns out this customer is
[20:55] larger and bigger and they are working in multiple segments and we have no idea about it, right? That's our green space. We can get more sales or we can approach the customer to look into the other areas because we do solve problems in
[21:11] other areas also. So this gives us a full picture of our you know, of our customer. The last one is just missing relevant information, right? We have wrong addresses. We
[21:29] We have I think one of the big things is the revenue. If you come to know about You know a lot about the the customer, but if you know how big this customer is. So in Dun & Bradstreet you can see how much
[21:45] sales was made. Like what is the revenue of this company, which is pretty public for most of the companies, right? And but you have to go and search and look, but here when you look at your customer profile, you can see the revenue and we can target the customers who are big based
[22:03] on the revenue, right? So that's what we were, you know, we we are kind of using D&B for right now. And what we did is we took our golden records and we matched it to the D&B
[22:19] information and our match rate was almost 80%. So we could find D&B information for almost 80% of our golden records. So which I think is is pretty good. The
[22:36] other 20% we have to figure out what to do. We'll get to that We'll get to that. We'll You'll take care of it. You bet you. But that that's actually There probably is probably because this is just like we just did this recently more recently so
[22:53] we haven't really spent more time on looking into what that 20% is. But we are happy with the 80% because we didn't have anything. So um, that's where we are. Um, so now I'm going to pass it on uh to Liz
[23:08] to talk a little bit about Dun & Bradstreet. I want to make sure this is forward. The laptop one? Uh, right one. Yeah, that's the one. So, let's see. Okay, so there's nothing
[23:23] better than to have someone like Madhu who helps us tell the story, right? Five different times she's done data management, master data projects. Many times when she's been in the commercial space, you've used master data and then used Dun & Bradstreet to support you in
[23:38] that. So, for us, we're going to tell a little bit more story about what it is she's getting and how it's supporting her. So, one thing she mentioned to you is it's about the end use cases, right? Not only is she doing her master data capability, but it's all about making
[23:55] sure that she's helping her business grow, she's mitigating risk, she's being very careful about compliance and making sure that this can scale to an enterprise-wide. You know, she can do that by herself, right? She doesn't necessarily need a third-party referential source to help her. However,
[24:11] that's just the inside view, right? She talked to you about hierarchies and the fact that, you know, she doesn't know how Shell might be aligned properly. So, using a third party brings all of a sudden that additional outside-in view into her business. It's absolutely
[24:26] critical. And she could pick any partner, but what is she doing? She's picking Dun & Bradstreet. And why? It's because our Dun & Bradstreet commercial graph. There's nothing else like it out there. It's centered around the Duns number, our proprietary um, external key for
[24:44] you. Um, and think about it from the perspective of this commercial graph. It's We're trying to identify every single business out there, so that when Madhu wants to understand and gain some additional insights about that business, she's able to get it right from Dun & Bradstreet. 650 million, sorry, 640
[25:03] million records, entities identified, over 12,000 data points. We have contacts as well, and linkage across the board here. But that's not the really cool part. We do 100 billion data checks, quality
[25:19] data checks every single month. We actually have over 1 billion API transactions every single day. And when you look at that compared to say major credit card company, they do about half that. So we have become a really
[25:36] critical component to data infrastructures across the board with businesses. And and really for us, it's about who's using it. 2 million users are using this information. It's absolutely critical that you're setting yourselves up for a foundation
[25:52] of truth, and that's what our commercial graph is all about. So if you're going to be using AI, if you're creating your master data capability, you don't have to worry about is this right? Is it wrong? Because that's our job. That's what happens when you add us to your data ecosystem.
[26:08] And then last but not least, you heard from Madhu how she's using it, right? We partner with Lake Fusion. So while they're doing matching, our job is to match to the DUNS number, to bring all that together. And we match it, we enrich the records, we have those hierarchies, we make sure that we're
[26:24] setting that foundation of truth and that trust. You know, this is not just about broad business coverage. This is about a trusted business identity and contacts layer so that when agents need to go and use this data,
[26:39] it's actually trusted data. There's no worry about it. That's what we're trying to bring to Madhu and to Lake Fusion. We are the third leg of the stool. You definitely don't want to do this alone. We're basically trusted, verified, and AI ready. But, don't take my word for
[26:56] it. Medusa will take that work on. That that that leads me to a very interesting point. Um So, talking about this hierarchy, um we
[27:11] we have our own hierarchy, and probably many of you know this, right? The company thinks like this is the hierarchy. We own it. We have our own hierarchy. But, then now we're showing them what the reality of things is. And um
[27:28] it is it is not easily acceptable, right? For them to change what they thought was the hierarchy. But, mergers and acquisitions happen, right? And that is one of the biggest things
[27:44] that we do not know what the company was is now laid, you know, completely, like uh I don't know, uh Salesforce bought Informatica, right? So, otherwise we would always think Informatica was separate, right? So, things like that
[27:59] will always happen, and we have to be ready to you know, uh combine all of that. Um So, ultimately, how are we using it? We have a view, we you know, we can profile the
[28:14] data. We can resolve and match the discrepancies early, right? Um So, so basically, I'm I'm not going to read all of this, but there is a lot of usage for your master data, right? Not
[28:32] just um uh not just providing it to other systems, but also cleansing it and keeping it real. So, challenges, nothing comes without challenges.
[28:47] Um technology is the easiest part, especially for a master data management project, right? The technology matches it, but then the data stewardship is one of the biggest failures of a master data
[29:02] management project. Um this has been talked about for years and years now, and I've been hearing this like forever. Um we also struggle. I I I mentioned earlier like when we started, right? Like we started in August, and I think
[29:19] we were done in December. Right? And from January to almost April, we in IT were shuttled from team to team to team
[29:34] to clean that 20% or not even 10 uh sorry, 10% that we had. Right? So, that is you know, even though we had engaged the users, even though I had that experience of
[29:49] you know, the data stewards not being available, I knew this was going to happen, and we engaged them from the very beginning, but in the end no one wanted to clean the data. So, that is something that we have to be very, very vigilant about if
[30:05] you want this to be a Finally, we have data stewards, we've created a forum, and the management has agreed because it's important for the management, right? So, they have made it happen now. Um the rules, it needed a lot of discussion
[30:21] because the customer doesn't belong to one team. It's just not finance's issue. It's sales, finance, compliance. Every team cares about the customer. So, no one really owns it, but everybody wants it, right? That That's That's the story of the customer. And so,
[30:37] everyone has their own rules, right? I want to do it this way, and I think this is what it So, that takes a lot of time and effort and you know, bringing people together. So, that's one. And then, adoption is
[30:55] I gave you the example of the hierarchy. What Dun & Bradstreet is giving us is true mergers, acquisitions, true hierarchies. And here we're saying, "No, no, no, this is separate." Right? So, that adoption, that mindset that you
[31:11] have to change to say, "Hey, this is reality." Right? We were living in that world for many years thinking this is how it was. So, all of these are like true challenges. And I don't even want to name the number of
[31:27] systems we have and how messy the data it was. And it still is, but at least for reporting, for the future, we have it. And now we are sending it to, you know, some of the systems that we are implementing.
[31:43] So, at least the future is going to be clean. Right? Uh we still haven't decided whether we're going to Everyone now thinks like looking at the data, they're like, "Oh, can we send it back to the source systems?" But that's that's a huge undertaking. I mean, that's a huge
[31:59] challenge, a huge undertaking. Um it'll happen eventually. Probably like, you know, look at the data that you want to fix, go and so fix it in the source system. But that that is a true challenge, especially when we have so many systems, source systems.
[32:16] So, um having said that, what does our future look like? Right? So, like I said, we've already started mastering uh our sites, um customer sites. We we call them sites. This is where we have all of our
[32:31] installations of our equipment. Um and then we are working on products and suppliers. We'll start with suppliers, actually, and then go to product because products is a huge mess, like, you know, and it's it's a lot of data. Um and there's no clear definition of
[32:47] what is what. Um but that's a normal challenge, right? Um of course, the market segmentation is going to be huge for us because that is where, I think, we will have more sales because we really want to tap into because, like I said earlier, ours is a
[33:04] niche market. Right? So, our customers also are very specific. So, we can't just find customers everywhere, right? They're very specific. We got to look for the specific customers, and so unless you have that additional information about
[33:20] our customers, we're not going to solve that problem, right? And um and of course, AI needs what? We are all talking about AI needs clean data, right? So, you throw in the garbage, and
[33:36] then the garbage comes out, right? So, for us to enable us to the future, this is, you know, this is a continuous journey. It's it's not a, like, you know, uh we're not done. And this is just the beginning, right? I'm super
[33:52] excited, and I always talk about it because um like I said, I am passionate about data, and I know everyone wants to go to a but you know, use AI and do lots of fancy things with it, but unless your data is clean
[34:10] and pristine, you cannot get good answers, right? Um we are using Databricks uh very extensively. We are building apps on Databricks. Um and so, you know, this, especially Lake Fusion,
[34:27] being part of Databricks, is, you know, solving our one-stop solution. Um Dun & Bradstreet definitely is um is helping us with additional data. We actually I forgot like we we also use Dun & Bradstreet to check on our
[34:43] our uh do credit checks on our customers. Um uh because you know, like I said, our what we manufacture is huge. It's like multi-million dollar projects. So, we cannot just start
[34:59] working with a customer without doing that. So, we use Dun & Bradstreet for that also. So, I think that's about it. Any questions? We're all ready for 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.