Genie for Enterprise Finance: Data Adoption and Governance at Scale

Summary

AkzoNobel, a paint and coatings multinational with 10 billion in revenue and 30,000+ employees operating across 150+ countries, transformed data from a liability—with 50+ ERP systems and answers taking weeks—into a strategic asset by building a Databricks lakehouse foundation and adopting Genie's natural language interface.
The team identified five keys to Genie adoption at enterprise scale: establishing a solid data foundation first, securing an executive champion, creating IT-business collaboration contracts, building trust through governance and SLAs, and reskilling analysts to work with the new interface.
After the implementation, AkzoNobel's CIO can ask finance questions directly from his phone during his commute and receive answers in seconds, a capability that previously required an analyst to manually compile data over days.

Genie for Enterprise Finance: Data Adoption and Governance at Scale

Enterprise data is either an asset or a liability. With 50+ ERP systems, 200+ legacy systems, and hundreds of dashboards, AkzoNobel faced a critical problem: business questions took weeks or months to answer. Watch how a multinational with 30,000+ employees transformed data into a strategic asset using Databricks lakehouse architecture and Genie's natural language query interface.
Discover the five-step approach to Genie adoption at enterprise scale: foundation first through ERP consolidation and data harmonization, finding your executive champion, establishing IT-business collaboration contracts, building trust through governance and SLAs, and reskilling your analysts. Learn how Genie answered finance questions in seconds that previously took days, and see practical strategies for production-grade governance, ownership, and agentic capabilities for enterprise teams.
🤝

Chapters

FAQs

What data challenges did AkzoNobel face before adopting Genie?

AkzoNobel operated with more than 50 ERP systems, over 200 legacy systems, and hundreds of dashboards, making data a liability rather than an asset. Business questions often took weeks or months to answer because analysts had to manually navigate fragmented systems to compile results.

What are the five keys to Genie adoption at enterprise scale?

AkzoNobel identified five keys: building a solid ERP consolidation and data harmonization foundation first, finding an executive champion to drive organizational change, establishing formal IT-business collaboration contracts, earning trust through production-grade governance and SLAs, and reskilling analysts to work effectively with Genie's natural language interface.

How did Genie change how AkzoNobel's leadership accesses data?

After the implementation, AkzoNobel's CIO can query company data directly from his phone during his commute, asking finance questions and receiving answers in seconds. Previously those same questions required an analyst to manually pull and process data, often taking days to complete.

Why was reskilling analysts a critical part of AkzoNobel's Genie adoption strategy?

AkzoNobel found that analysts needed to shift from writing SQL queries and building dashboards to curating data context and validating Genie outputs. Without intentional reskilling, analysts risked becoming obstacles to self-service adoption rather than enablers, since the value of Genie depends on well-governed, well-contextualized data.

Full transcript

[00:09] Welcome everyone. How fantastic that you guys are all here after three, maybe four, maybe even five days of this conference. Um, we're super excited to be able to speak to you guys today. Um, and thanks for joining us. So, I want to start out with a quick
[00:25] question because, well, I also want to hear from you guys. Um, what does your CIO do on his commute to work?
[00:40] Anybody? This is a real question. I really want somebody to answer it. So, preferably uh one or two people. How many people? Sorry. How many priority one tickets for today? Okay, he's asking that question and he's he's calling he's on the phone. Okay, he's asking that question to probably an
[00:56] analyst or something. Awesome. Anybody else? Maybe they're reading the news or listening to a podcast or maybe watching some Netflix or something. Um because of course our CIO, he actually
[01:16] looks on his phone and he actually asks the questions about maybe Jerro tickets or something else right inside our own data. And that is the journey that we're about to tell you today. how you can also get your business to use Genie in the correct way. We're really excited to
[01:33] bring you along on this journey. Unfortunately, this talk is not really technical. It's really about how we got the business to use Genie. Hi, I'm Christy Blazdall, team lead of enterprise reporting within uh Oxonabel.
[01:49] I'm super excited to be back in my home country. I'm originally from Chicago if you can't hear from the accent, but I have lived in the Netherlands now for 10 years. Paton is going to talk to you about how we created the foundation and that was the start of the the born of of
[02:06] Genie. I am going to tell you how we actually got the business to use it. And a small secret, my part was the hardest. Fion over to you. Okay. Thank you very much Christie and good morning and welcome also from my side. My name is Paton Snoop the main
[02:23] architect within a debell and uh I'm from the Netherlands I live in Amsterdam. So it's not only a multinational but we're also a multinational presentation that we're about to give you about let's say the adoption of Genie in our organization. But before there's let's go first a
[02:41] little bit to our organization. So, Axenel, you might have heard of us, you might not. Uh, we are a painter cotics organization and we operate, let's say, in in over 150 countries and markets. We have factories localized in 124
[02:59] countries, maybe even 20 126 at the moment, 10 billion uh revenue, and we employ over 30K employees. And that number is important later in the presentation. So we have brands like Sickens, uh, Flexa, uh, Lasol and you
[03:16] might find us on different surfaces. It might be a train, might be an an air, uh, craft. Uh, you might find this on your car with our coating. Uh, you might find this on buildings like lasagraphia, which is a big cathedral in uh, in
[03:34] Barcelona in Europe. the London Sharf which is a tall building over there. The the Sydney Harbor Bridge, those are iconic iconic locations. The Singapore uh Bay uh gardens also contains our
[03:49] coatings. But also in this room, most likely multiple objects in this room are coated with our coatings. It might be the chair, it might be the uh the the the wall uh panels, it might be the ceiling. Um, most likely our coating
[04:07] will sit on your phone that you use, your mobile phone, might be the can of soda. Basically, you can find us all over the world, but most of the time we are just visible as a coating on a product. Maybe two places where you might not expect us directly. We are
[04:23] also in your bedroom most likely as paint on the wall only to paint. And you might also find this on the Mars planet where the rover that is running around to collect let's say materials for for for scientific research. That rover that
[04:38] is running around there is coated with our coatings. So you can imagine you can find this in different places uh also very close to your home but also very far away. Um if you have such an organization with so many different brands, so many different
[04:55] markets, but also so many factories and so many employees, data it is either an asset or a liability. And in our situation, I can tell you already at the beginning of our journey, it was
[05:11] liability. However, we turned into an asset. There's no middle ground here. It's one of the two. So where did we started with our let's say foundational our journey towards the
[05:26] success the genie is bringing into the organization at the moment we have to start with the problem and the problem was severe the problem was big and the problem is a problem you might recognize within your own organization um so it might resonate with you but
[05:42] most likely you will at least recognize it and this was the problem. So when I started with Oxin Nobel which is already long time ago, we had over 50 ERP systems up and running in different parts of the world and in different
[05:59] business units which we applied. If you would also include the nitty-gritty ERP systems, sorry for the word, then we would have had over 200 ERP systems running in our organization. All those ERP systems were multiplied with different analytical solutions on
[06:15] top of that. So in order to get a enterprisewide view of any topic in the organization there was a lot of work to be done and what you see over here is that let's say Excel was really duct taping everything together. So you could think about finance but you could also
[06:32] think about supply chain or manufacturing. Every question that came from the organization was a problem because it was so scattered everywhere. So the problem was not only the technology however it was also our organization it
[06:49] was the definitions of the KPIs which he applied it it was also the processes which he applied we hadn't had any standard or structure whatsoever we are an organization grown through mergers and acquisitions and what you can imagine every emerger every acquisition
[07:05] or every organic grow that we do in a new country or a new market brings new technology new processes new data, new KPIs, new management that all think our definition is the best for the whole of Axon Bell. So good luck it to solve that. So this is the situation in which
[07:23] we were and to get any question out of the organization it took us more than months even sometimes half year to get the answer in a decent spreadsheet or a dashboard to our senior management. And guess what? the problem had long gone by
[07:40] then already because they would have found the information in a different way or did a simple guess there. So we understand we had to solve this problem if we really want to begin a global multinational with harmonized processes and also harmonized technology and data
[07:56] that could be turned into an asset. So this is the problem. What we did, we were adding dashboards to the problem. That was our answer to the
[08:11] problem when we started. We were building dashboards by the hundreds. We are having currently in our organization more dashboards as employees. And when you remember 30,000 plus employees, that's a lot of dashboards in our organization.
[08:26] So we tried to solve that and we had many discussions with the business. We are from it. So our business was pressing us and pushing us as it they were smashing us around through the organization. Give us the insights and the information
[08:42] quicker. So it guess what step out we will take over as business. We will start selfservice analytics. So after long discussions political but also technical but also organizational we allowed to open up our IT landscape
[08:59] the reporting let's say in the desk morninging part there to go into more selfservice analytics. This was already after the moment that we did our first attempt to harmonize and to globalize by centralizing the IT
[09:15] divisions into one central IT organization. Well, guess what was the first thing that the business did when they get access to the development tools to develop their own reports and dashboards? They outsourced it. So, they went to
[09:30] third party boutique vendors and they outsourced the sales service that we provided them with to deliver their dashboards by this third party vendor. This did not solve the problem. It only shifted the problem. It increased the problem because those guess what those
[09:47] third parts they do a great job they make fantastic dashboards but they make it so complex that the business could not run them anymore and then they would come back to it listen this problem here with this dashboard sounds like it to me solve it please sorry it is self-service
[10:04] no it so we were having so many discussions in organizations we could not uh stop that So we knew we had to do something bigger to overcome this. So what did we do? We started we realized
[10:20] we have to fix the foundation first to overcome this problem where we have so many dashboards and even with all those dashboards we were not able to answer a single question in the organization appropriately. So what did we do? We went for the
[10:38] simplification and this was a huge undertaking. I already told you we harmonized our IT organization into a centralized IT organization. That gave us the mandate as IT organization to also start harmonizing
[10:54] our ERP landscape. We did it big. We went from plus 50 large ERP systems to now only four. What we also did, we harmonized all our finest data in a single central finest instance. This brought us a harmonized set of general
[11:12] ledger data that could be used across the organization. We went also from let's say the excels which I mentioned which are duct taping everything together. We went to a lakehouse architecture and we implemented data bricks. Why data bricks? Because at that time we could
[11:28] have made many many choices. We could have go to different software solutions or platforms. We selected datab bricks because of its openness. It integrates AI with analytics and also let's say the road map of data bricks was very
[11:44] promising at the time already. This was before genie was at the table even then it made us uh uh it gave us at least the possibility to derive insights and reports at a much faster rate as what we
[11:59] could do previously. But still if you look at the report or if you look at the dashboard you are looking behind you. You're not looking forward. It gives you a sort of static view what has happened in the business. It is not actionable.
[12:14] You cannot move the needle with a report or a dashboard. It is only the action that changes something in your organization. So that is the moment that we started to realize we have to go one step further. However, this was also the moment that
[12:33] all of a sudden datab bricks launched Genie. I think it was a year ago. It might be even be two years ago. I don't know exactly when the moment was. But basically Christy and myself when we realized Genie was there in the datab bricks platform, we got really
[12:48] enthusiastic because what we realized here we have a pot of gold sitting in our technology. Not yet in our organization but in our technology. It's called Genie. And this is the big promise that it gives you that you can go with your natural language into the
[13:05] data. So we were very untoastic about it. But that was the two of us. So when we saw this we understood hey this can be a gamecher. So
[13:22] we were tried to change the organization and to change the data from a liability to an asset and let's say from reporting and dashboarding into Genie. So we tried to go from a problem that was still there at the table. It was only shifting
[13:38] everywhere in the organization by everything that we did into an opportunity and this opportunity that was real. We realized that but basically we were let's say because we are of course let's say an IT organization we see data bricks before the business sees it. We were the first to see this and to
[13:55] realize this that here is an opportunity not a problem but an opportunity. So this is the vision that we put on paper. Nothing special, nothing fancy, nothing big. You want to know something, you
[14:12] ask, you get an answer. That's it. Nothing else. It doesn't matter if you're the CIO, the CEO, you're a business analyst in the organization. What you can do, you can
[14:28] just chat with your data and you get an answer based on the question that you ask. One of the problems which we had with the dashboards that we created is that we are an coatings organization. A lot of our work is about color. So we had
[14:43] hundreds of business analysts that were just beautifying dashboards. Give it the right color. Give it the right layout. Of course, it's very nice. It looks good in the eye, but in the end also a beautiful dashboard compared with a not
[14:58] so beautiful dashboard doesn't move the needle. The only thing that moves the needle is an action in the organization. So with Genie, it gave us the opportunity to step away from the content and the form and we could go into the context. That was the big
[15:15] change that we could apply within our organization with Genie. What we realized however the two of us having a good idea doesn't do anything. So we had to bring this vision and this technology that we consider to be a gamecher into the
[15:32] organization. And this is really where the hard part started. And this is where I would like Christie to take over. Thanks, Pat. I'm going to quickly grab a sip of water.
[15:50] All right. So, no story is complete without its challenges, of course. And we had two two that sat between us and our vision. But the greater the challenge was, the more the success was bigger than we ever
[16:07] could have imagined. So we needed to find somebody who was going to be an advocate, somebody who was going to actually put Genie to the test. So we built some dashboards or we built some demos for uh for Genie and we
[16:24] brought it to the business. But literally all we heard was why do I need Genie? I can just look to my dashboard. Little do we know that actually the questions they were asking were not the right questions. I'll talk to you about that in a little bit about the mind
[16:40] shift that we had to make. Have you ever had that feeling where your head is in the sand and you're screaming, "This is amazing." But literally no one can hear you. That's exactly how we felt. Like Patreon
[16:57] already said, we really saw Genie as a gamecher. Why is no one in the business wanting this? You can literally ask anything, anything to your data. Why wouldn't you want that? So, we finally
[17:12] sat down. We said, "Okay, we need to change something. We need to do this differently. We need to go to someone else. We need to find this champion or this advocate." So, we're in it. why not go to our CIO? We decided to change it up. Instead of
[17:28] creating a just standard demo, we actually went to our finance data, the data that sits in front of our CIO basically every week, if not every day. We went to the business analysts and asked them, "Hey, what are some
[17:44] questions, really good questions, that you're having a difficult time answering, but that our CIO has?" We let Genie answer them. We created a really good demo and sat together with our CIO Patton and myself and a few other people
[18:00] as well as our sales rep from data bricks and we showed them we said look these are the questions you've been asking the business analysts that maybe they take a bit of time to understand or actually get the information but here it is within seconds
[18:18] jaws dropped. It was amazing to see like literally everyone went quiet. But we decided to do something else in that session which I have to say was a little bit a little bit scary for Pon and I. We decided to go off script. We
[18:35] decided to ask our CIO. What is the one burning question that you've been wanting to ask a business analyst but don't dare to because it might be a tough question. Give it to us. Let's ask Genie. Of course, keep in mind that this was on the finance data. So, we made
[18:51] sure that he knew the data set and understood what Genie could edit Jeanie answered it. Genie literally within seconds was able to answer the
[19:07] question that was on the top of the CIO's mind. And this is where also our CIO finally saw Genie as a ch game changer. It was no longer a tool that he could just play with, but literally a means to actually put it on the map to
[19:23] give the business actual insights and actionable insights. So, of course, like I said, we had two problems. One was, of course, on the business side, so making sure that
[19:38] people understood what we were doing. But platforms change a company people do. So of course we all have a trust issue, right? A lot of us have dashboards that have been built for a long time. People have seen those
[19:53] dashboards and they really trust them. We had to do the same thing to Genie. So every Genie space needs to have an owner, an SLA, and hopefully a certification stamp. I have to say when Pion and I were on
[20:09] our journey with Genie, we had um we had used our our Steven data which is a massive data set. It's around 10 billion uh rows and counting as well as 50 500 columns wide. I wish it was only 50 columns. Um
[20:27] but when we were charting Genie, uh we actually had a PowerBI dashboard next to it so we could make sure that the numbers were correct. So one day I was asking Genie, I said, "Hey, give me the AITA for a certain business unit." Unfortunately, Genie
[20:45] failed. Genie gave me the wrong AITA number. And it wasn't slightly off because, okay, let's be honest, Genie is still an LLM, right? So it might hallucinate here and there. Your numbers aren't always going to be 100%. That's okay. However, this number was totally
[21:01] off and I was like, "Oh no, how are we going to change this?" So, I went into Genie. I thought, "Okay, what is the calculation incorrect? Is something is it is it getting context incorrect?" Then I thought, "All right, what if I just took a screenshot of PowerBI
[21:17] dashboard and gave it to Genie and simply asked it the question of what's the difference?" And guess what? it was able to tell me exactly what the difference was. I have goosebumps even thinking about it. How cool is that?
[21:34] Literally a snapshot. It wasn't that I gave the DAX query. I didn't give any information or background information or anything. I just gave it a a screenshot. Then I went to the finance department. I said, "Hey, listen. These are the two, you know, uh uh um uh things that I have
[21:50] in the background. Which one's correct?" They told me I needed to change Genie. Now within 5 minutes I had changed Genie and it was correct. That's when I knew we had to make sure that Genie was trustworthy. One of the biggest things that you also need to do.
[22:08] The next is of course reskill not rescale not replace. Sorry that doesn't make sense. Not reskill and not replace. Your business analysts are probably hired with the thought in mind that they're going to actually analyze the
[22:25] data. But what do they probably do? Build you dashboards, right? However, they have so much understanding of how to calculate things within the business. They have business knowledge. They understand the requirements from your business from
[22:41] your business teams. They have so much understanding of that and that is truly powerful. They can bring that knowledge to Genie. Ask the questions directly to Genie. Why not?
[23:01] Yeah. A new contract. So, of course, we all like to say we like to collaborate with it and the business. Sometimes it doesn't always go great, but here with Genie, it's really what is needed. And I don't know about you guys, but this past three days, one of the best things that
[23:16] I've heard is about Genieontology. We cannot wait for this. We cannot wait to get our hands on it and get home and actually start trying it because this is really what we've been missing. Of course, that connection between business and it to really give the context to
[23:32] Genie. Of course, we have metadata and we try to put in as many as many instructions as possible into Genie also with metric views, but that's always that's never enough, right? Finance has tons of SharePoint forms. They have all
[23:49] types of of things in every single uh uh sense of the way of Excel spreadsheets that also Genie could use as context. So, I'm excited and I hope you are too.
[24:04] Now, sorry. The moment of truth. We've had our two challenges and we've tried to overcome them. But now it's time for an actual project. So, you've heard us talk about the finance project. That was one of our biggest projects the past year,
[24:21] the past sorry, two and a half years that we've been focusing on. Now, of course, Genie was only born about a year ago, but for the past two and a half years, we've been fully focusing on the foundation. We've moved all of our CEO data, which is SAP based, into data
[24:37] bricks, making sure that we've built the balance sheet, the P&L, margin management, you name it, anything in finance, we probably have it. Then when Genie came, we were able to put that exact Genie on top of all of those models and actually create value.
[24:55] So today we're a able to ask not only strategic questions but also those nitpicky questions. For example, how much do we spend on laptop bags? Would you actually ask your analyst to
[25:10] go analyze that? Probably not. But Genie, why not? Now, of course, on the left side, we have our traditional dashboard. I'm not saying that we're going to get rid of dashboards and never see them again. Of
[25:26] course not. We always want a grounding point for everyone to see the same numbers and the same thing. But why not also be able to ask different questions and other things to Genie.
[25:43] I'm now going to take us through a smart a small demo. It's not on our data. It's synthetic data that we built together with data bricks and it's a very very very small data set. So bear with me while uh the uh that it goes through. Um it's about churn. So churn is basically
[26:00] what customers are leaving your or potential customers are leaving your organization. Um, yeah, I will.
[26:16] Yeah. So, I'm a sales director. Give me my insights and customer turn. So, here you can see JD has uh uh produced an overview perview given me an overview of what I what I what I have within my data.
[26:32] Awesome. And it's always also recommended some actions. But I want to go into more details and I want it to give me a plan about our degrowing customers. So here you can see that it's actually given a plan, a very detailed plan of how I can do this with root
[26:49] cause analysis, an overview of the combination of revenue losses and what specific customers I need to go to and a priority auction plan. And the funny thing is is you don't necessarily need to say, "Yeah, but
[27:05] this, that, and the other isn't correct. It's the signal and the noise that makes the difference." I don't know about you guys, but sometimes I have a difficult time just starting things, right? So, if I have a blank piece of paper in front of me, I think, "Oh gosh, where do I
[27:21] start?" But with this, with Genie, it allows you to not have a blank sheet of paper in front of you. Sure, maybe Genie will give you 10 actions and maybe you say they're all crap except maybe two of them. But hey, at least those are two
[27:36] actions that you have that you didn't have before. Oh, sorry. I'm seeing myself myself speak. Yeah. So, what's next? Everything that I've said, of course, is something that we've
[27:53] already done, but of course, there's so much more. The one thing that we would like to do in the comingings and the coming months, years is uh proactivity. Of course, one of Patreon's pet pep
[28:08] projects as we like to call it um is listening to data. So like the first question that I asked you guys was, what does your CIO do on their commute to work? Of course, we're from Amsterdam, so we all like to take the train. However, we're also from Amsterdam and
[28:23] everyone likes to ride on their bike. So what if we could actually listen to data? What if instead of having Genie be in front of us that we have to read it, what if we could listen to it? That would be pretty cool, right? Contextual.
[28:39] So at the moment when you have a dashboard, for example, you have to give each individual person access to it. But what if you didn't need to do that anymore? What if when I came into a company, when I started, I would have
[28:54] all of my information right at my fingertips? Wouldn't that help you? This is something that we hope that we can really achieve in within Oxonobel to make sure that everyone immediately has access to the data and the genie spaces or the AI dashboards that they need to.
[29:12] And of course, Agentic. We want to make sure that people have actions and can take actions on those things that Genie gives them. Before we close off, I want to tell a small story. So, our sales d our sales director was
[29:28] one of the people that our CIO went to. However, of course, he said, "Yeah, but I'm not going to have my salespeople be told by an AI agent what to do." We said, "Okay, totally understandable. But let us or hear us out. We're going
[29:46] to show you a demo on his own data with his own information and questions that re he really has." Within a few weeks, we were able to convince him that indeed Genie is a gamecher and he can actually help his sales reps get more insight
[30:03] into what's going on. That turn that you just saw, that was a real problem that we actually took into Action Nobel. We went to sales director and showed and said, "Hey, but do you recognize these clients?" They immediately recognize them.
[30:18] Why not use Genie? And of course, the production maturity. Yeah, we're not there yet. We're still working on it. Now, I know some of you might be tired from yesterday's big party. Maybe you had one too many beers, or maybe you're
[30:35] just super excited for the week. But if there's nothing that you've remembered from the past 20 minutes of our speech, remember these five things. Fix the foundation first. The last two and a half years, we
[30:51] focused on getting the finance data into our data bricks. We made sure that everything was quite literally perfect. in order to put Genie on top of it. That was really what made a difference.
[31:09] Win one executive's heart. I think if we hadn't done this and we hadn't gone straight to the CIO, I don't think Genie would be as big as it is today. And the funny thing is, if you look at all of the information behind Genie, our CIO is the top user of Genie. He uses it
[31:25] literally every day. So the fact that there's now an app and whatever else is to come, he can't he cannot wait. Number three, write the IT business contract. Make sure that there's a collaboration between you as IT people
[31:42] and the business. Number four, treat trust like a product. Like I said, have those standards. Make sure that you trust the product because if you don't trust the product, business won't either.
[31:59] And number five, move people up and not out. Your business analysts are extremely important to the organization. They have the knowledge. Let them in on Genie.
[32:14] Now, thank you so much for joining us. We're now going to open up for questions. And if you don't feel comfortable asking a question out loud, we'll also be out outside uh for the for 20 minutes or so. Thank you.

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.