Name: A New Era for Retail AI: SAP BDC Connect for Snowflake
Uploaded: 2026-01-22T19:16:08.643Z
Duration: 48 min 2 s
Description: A New Era for Retail AI: SAP BDC Connect for Snowflake

Transcript for "A New Era for Retail AI: SAP BDC Connect for Snowflake": Alright. Well, as we have people trickling in here, just wanna thank everybody and remind you that as we go through this, then use that q and a section on the screen for any questions you have as we have our little discussion here today. So we'll give it just one more minute for folks to finish entering the room and then get started. Alright, Freddy. Let's go ahead and get started. So I wanna welcome everybody to this presentation from Hakkoda, an IBM company, and Snowflake, where today we're gonna talk about, you know, the new era for retail AI. And this conversation is focused on what's that value that SAP BDC Connect for Snowflake is going to bring into the retail industry. And this conversation, just to level set for everybody, is very much focused around what are the business use cases. This is not necessarily a technical conversation. This is very much focused on what's that business value? Why is this important to the business and organizations looking to looking to take on SBBDC and and have it impact and change, you know, how their organizations are operating and working today? So if we go on to the next slide, I would like to introduce myself. My name is Matt Florian. I am the head of the SAP hybrid cloud analytics for Hakkoda where, you know, I work with them and anything that we're dealing with Snowflake and and SAP coming together and working in that space, that falls into our our practice. And we have a dedicated practice just for this where we work with all of our deep expertise in Snowflake and all of our deep expertise at IBM and SAP and bring that together into a cohesive one team for our customers as we work on these big problems like we're like Freddy and I are gonna talk about today. So then I'll turn it over to my friend Freddy for introduction. Thank you, Matt. It's always a pleasure to partner with you and Hakkoda. Good morning, everyone. Good afternoon. My name is Freddy Guard. I'm part of the retail and consumer goods industry team here at Snowflake. I lead up our efforts with consumer goods, customers of ours. And my focus is mainly on driving business outcomes and generating value for our customers through their use of technology and data. So it's great to be partnering with Matt on this exciting announcement that we have. So before before we dive into the SAP conversation, I think it's important to talk a little bit about what does the AI implementation look like in the industry today. You know, if you think about the last twelve to twenty four months in the retail and consumer goods industry, we've seen a lot of great POCs come about. We've seen a lot of one off projects that have been successful in generating value. However, as you might have noticed in popular press, there are also lots of data points where it's been mentioned that it's been hard to get true business value out of several of these outcomes and the projects that have been driven in the last year. So that is true. It it is relatively challenging to set up enterprise AI and scale enterprise AI successfully unless you start putting in the right building blocks in place. And we are seeing organizations that are now really starting to take enterprise AI seriously. They are putting together at least three building blocks that I see as key enablers in this space. So number one, it starts with data. You know? So making sure your ability to leverage all relevant data that's within your four walls, that is in in your ecosystem, and your third party datasets that you have available to you? How do you leverage that data? How do you mobilize that data? And then how do you, most importantly, train your models on that data? Why do we focus so much on the data? Because we believe in our industry and most industries in general. All organizations at the end of the day will have access to the same set of models. The difference is gonna come from what data can you harness and how will you use that to train your models. And the SAP VDC connector that we're gonna talk about is a big enabler in this space because as we all know, ERP is house a lot of important information in our industry. And SAP is the ERP of choice for most of our customers. So a lot of that data is within SAP. How do you make it easy to access the data and the semantic logic that's with that data and use it with other data sources that are non SAP? So that's the data bit of it. And we'll talk more and get into the details of the SAP BDC connector. The second big one is your data stack, your data platform, if you will. What data platform do you use? How do you use it? How do you really enable all of that data that you've collected to really start mobilizing that data in your organization to provide insights? We'll we'll touch upon that in a little more detail as well. And then the third big enabler, which I think is sometimes overlooked in a in a lot of these efforts, is the right governance on use cases and the focus on business value creation. So that means identifying the right use cases that are going to move your top line or your bottom line results and where the results of AI efforts that you're putting in can enable business users to fundamentally either rethink outcomes and processes or get deeper insights into their decision making process. If you're not starting with the right use case prioritization with this in mind, you are, I'm afraid, going to fall a little bit short in terms of realizing the right right value. So having said that, let's talk a little bit about what what is the data platform, what does that look like. So I will I will talk about it in in terms of the Snowflake data platform and what we are enabling in this space. So for those of you that are not very familiar with Snowflake, Snowflake is the AI data cloud company. We are the data platform that is ready for AI and ML. Where we really differentiate ourselves from others in the industry are on the three principles of easy, connected, and trusted. What does that mean? We truly believe that a easy to use platform is the best when it comes to scaling. So we are a managed platform. We can handle all data types, structured, unstructured, semi structured data on the same platform with all your workloads. The unstructured data piece is extremely important, especially as you think about enabling AI in your organizations. In our industry, we have a lot of unstructured data that is very poorly utilized today. The ability to enable utilization of that data will drive deeper insights for you. The second bit on the platform that we believe is truly a differentiator is the connectedness with our open architecture. It takes, it comes into shape in a couple of different forms. One, we are cross cloud, so we work across all three major public clouds, AWS, Microsoft, and Google. We also have an open architecture, which has enabled an extremely comprehensive ecosystem of partners, including Hakkoda, who have really started developing the right data applications for us and the ability for our customers to take advantage of things that have already been built for them, especially industry specific use cases. And then the third bit is trusted. You know, you need to be able to run all these large models in your secure environment on your data. And you need to be able to have enterprise, you know, disaster recovery features. You need to be able to have cost per cost governance features, especially when it comes to running AI models and making sure that you have complete visibility and control in a secure manner as you run your models on the data that you have. So with that being said, when we let me introduce the SAP BTC connector. We truly believe that this is going to make it much more simpler for our customers to take advantage of all of their SAP data, merge it with non SAP data, and really start driving deeper insights as you start activating your AI use cases. Over to you, Matt, to get into some of the details. Yeah. Excuse me. So this this connector, this announcement that came out here recently is really in the in the world of being AI ready, having a foundation for AI, this becomes a a game changer for how we're working with that data between SAP and Snowflake. The overall objective is that we're taking away the friction, making it easier to work with while maintaining that strong governance, that strong security model that SAP is known for. And then sharing of this isn't about you know, traditionally, we've had the issues about how do I get that data out? How do I move that data? And it's the conversation's always been one directional. How do I get that data out of SAP into Snowflake and that movement? With the SBBDC Connect, what we're talking about now is how do I share that data? I don't have to move it. I have to I'm sharing it because I'm sharing the semantic definitions and that underlying technical infrastructure that allows SAP and Snowflake to share bidirectionally. This means that for organizations, as they are embarking on their s four journeys and as they're building out their intelligent applications inside of SAP that you it really helps keep SAP clean core and activates, you know, how they're operating and how they're looking for those operational efficiencies that the two coming together. And as as Freddy talked about that platform, choosing the right platform, now it's not I have to choose one platform. This is about choosing the right platform for the right work where SAP excels at SAP related work, and Snowflake excels at bringing together the rest the enterprise data and bringing it all together as a cohesive enterprise strategy, which is what we need as the foundation to succeed for AI use cases as we build along. And that's what makes this such a a powerful combination that comes together. We're removing the friction, removing the piece that is been a blockage for these projects in the past and simplifying how they share and when they share. So if we go to the next slide, right, when you look at the market and you hear this being announced, you're gonna hear two different products being announced. You're gonna hear SAP Snowflake, which is a solution extension from SAP and means that if you don't have Snowflake today, you can get that through your SAP contract and have it as part of your your entire solution portfolio from SAP. And then there's SAP BDC Connect for those that already have Snowflake. The key for each of these products, though, is that the solution, be it SAP Snowflake or SAP BDC Connect, means that you're getting the full power of Snowflake. There's no removing of features that make this work. It's Snowflake, full times. So all the features and as Freddy said, our ability to work across the different the different platforms. That can be AWS and Azure, Google Cloud. Being able to cross and really have that strong enterprise platform is there. No matter which path we choose, it's there. So that makes this a very exciting announcement for how this comes and how this is going to really democratize data and open up the entire landscape to take full advantage and get full value out of your SAP investments as well as the investments that you made in the rest of your ecosystem. So when we look at this and we say, okay. Why is this important? And there's, you know, plenty of technical documentation, plenty of technical blogs have been around and said, what is awesome about this? Why is this great? Then you this comes down to the use cases. And with everything in AI and as Freddy, you talked about it at the very beginning slide, and we're talking about why things fail. It comes down to use cases and having the right set of use cases in here. So, you know, let's start diving into some of those use cases, what this is. And and, Freddy, you're at the front of this on the on the manufactured and, you know, consumer packaged goods and retailing. What does this look like to you? What what are the exciting things coming out of Snowflake for why this is important? Yeah. So like Matt Matt mentioned, we already have several 100 of our customers bringing SAP data into Snowflake to be able to add to other non SAP data to really just start generating insights. But going back to what to the to the earlier comment I made, right, the business use cases that really are going to be impactful as you look to generate scale with enterprise AI are going to be ones that impact your p and l. And there are a whole host of operational and functional use cases that have a dependency on SAP data And also the semantic logic that lives within SAP. Right? So if you're able to combine SAP data with other non SAP data, you are able to drive deeper insights. And this is just a, you know, a glimpse at some of the kinds of use cases that we either have customers implementing or are in the process of evaluating and thinking about next steps on this. Now this is by no means an all exhaustive list. And you might have additional priorities. You might have slightly different priorities. But think about it in the in terms of what is truly going to be impactful for your organization in terms of being able to really either generate more revenue or free up more cash for you with efficiencies. And those are the kinds of use cases that we really think are really going to make a difference. And and, let's say, yeah. really quick. I just in pointing this out, I think one the things that for those in the audience that are on the s four transformation teams and have been in these projects, When you look at these lists that Freddy's talking about too, right, these are all things that are part of those transformation programs. These are all objectives that we've seen in there, and we've a lot of times have been done very manually to how do I go and implement this, make this better. And these are the things that I think resonate should resonate, I believe, with most everybody in here. Yeah. And I think, Matt, like you and I were discussing just before we hopped on. Right? Things like tariffs and ever changing supply chain situations and disruptions that most of our customers are going to continue to face through this year. You know, the ability to react and the ability to plan ahead is what is going to really start differentiating, you know, business decision making in this next year because the uncertainty is going to remain. I think the other bit that's also important to keep in mind is our platform not only allows you to bring that SAP data once you generate the right insights out of your AIML efforts. The ability to package them into data applications and data products is also extremely important because then you can activate it with your business users without them having to worry about really understanding, you know, the technical details of how kind of you generated the answer and all the mechanics that went behind it. So let let's make it a little bit more real, Matt, by diving into a couple of these examples and and exploring that further. Yeah. So let's let's double click in a few of these. Right? So we're gonna talk about agentic allocation, something that happens often for that order to cash scenarios. We're gonna look at live logistics and how this can help with that as well as customer three sixty. So let's first dive into this agentic allocation. So then the allocation of of available inventory of incoming purchase orders to open orders and how to fulfill. I mean, this is a common problem that that every retail and every consumer packaged good company encounters. Where do I go and how do I go and and make sure I get those quick cycles on my inventory and get the customers that are my priority to get them their products and get all those open orders and meet my commitments. Historically, these kind of programs have been very, very rigid and rules driven in what they've done. Sort my my open orders by this criteria. Sort my incoming products by this criteria. And it's been, you know, a very human intensive activity that goes on to what do I do and how do I make this? And not always the most optimal decision, but even when it happens and I go and make a decision on what the allocation is, then it's, you know, handing over to another team to say, alright. Go back into SAP. Go and and do these allocations and watch throughout the day as your allocations tick and dip down and you go, alright. I met my plan, and it took me a whole day to get to my plan to make that happen. One of the areas that we see this this capability that comes into play is this being able to do that allocation, but using this agentic framework to go and and activate it and and take out that human entry element that doesn't necessarily add the value. It's the allocation that adds value into the work stream and gets this this moving faster as well as reduce errors along the way. So in this kind of a use case, right, we're gonna see activity happen along three main ways. It is, first, being able to drive value by using your stronger, better, more optimized machine learning that would go and do this allocation using AI and machine learning together that says, here's a more optimal. Here's these are our core customers. These are the customers we have big commitments to. These are the customers that we're growing. And being able to take that SAP data with, say, data from Salesforce and data from sitting over in your warehouse management systems and any of these other pieces coming together saying, alright. This is the inputs that I want. What is my optimal revenue potential for allocation? And once that decision's made, once the recommendations are there, with that human in the loop to go and improve the recommendations or say, refine it, improve it this way, the human in the loop that goes and tries to optimize that a little bit further and makes the decision saying, this is the right answer, then upon that decision making, say, allocate. And that allocation then executes the agents back in SAP to go and create the products and create the documents that allocates product to open orders and then really activates that supply chain to move that product to that customer. That shortens the cycle for order to cash. That shortens up how quickly we can go and start allocating and reduce errors that along the way of allocating the wrong thing. So it really speeds that delivery and speeds the decision making process, but keeping humans still in the loop to make the final decision on what's there. I think this is one of those big use cases that goes and then looks at that how that data is flowing in both directions, which is what really is critical in this, you know, ability with SAP and and Snowflake and BDC Connect acting as a conduit that makes this data flow in both directions. That's a big differentiator here. So the next one, if you look at the next piece in here, and it would be to transform supply chain. Right? So there's always this ability and always this desire on how much more can we squeeze in the supply chain to be more predictive instead of reacting to things. And we've seen this the use cases for this time and time again where there's supply chain disruptions. It could be disruptions because of some event in the other side of the globe. It could be disruptions because of a weather event that happens. You know, any of these kind of situations, that's something that affects supply chain and and how we can go and be more responsive to that and and act ahead of time on that supply chain to keep your the flows of goods moving quicker, faster, better, that gives advantage. So this is, again, one of those areas that, you know, we were looking at data and coming in and using SAP and using our orders and using the data that we have about what those open orders are, what is in transit in there along with the data we have from our inventory management, data we may have from an external warehouse management or from a third party logistics provider, from weather, you know, all these different pieces that come into play that can go and help influence and be more predictive in this live logistics model. And this to your point, Freddy, on, you know, how does this affect with tariffs or this change in here? All these things that can impact that supply chain, you know, having a having this drive into predictive that uses both SAP and Snowflake. And, again, it's about this round trip of data. It's there's no longer are we looking at data as just this linear flow that goes and leaves SAP and never comes back. Right. SAP. B2C is about coming back. Absolutely, Matt. And I think we are we are already seeing, right, supply chain organizations that have been on a transformation journey for the last several years. You know, the the holy grail is, like, how do I get to better demand signals? How do I get, you know, better signal or proactive signals at any supply disruptions? And the ability to bring in third party data along with everything else that's happening within your four walls, which a lot of it is within SAP, is really starting to drive that complete landscape of information that then you can use to drive your decision making. Now your decision making is no longer siloed by systems of record that hold your data. You know? That's that's part of the reason we've had some of these challenges with supply chain in the past, and we are able to get past it now. Right. And it's you know, the when I've worked with SAP customers, Freddy, in in that supply chain space, one of the things that we often find is that they are dealing with what they have inside of SAP is a maybe it's a summary of what is coming in. But and and it's important because that summary and then the value, then how much quantity, what the dollar amount is, you know, all those key things around that are driving the cost of goods that are driving up, you know, and saying this is what our margins are gonna be. And that that core that really hits the the critical KPIs for any organization is important, But what was lost, what was missing in the decision making for operators and for the models that was coming in was there was so much other detail in these other systems that was not being taken into consideration when we're building these models. And that's why, you know, the SAP is important. It's I describe it to some of my, you know, folks as it it truly is the heart and circulatory system of the enterprise. But there are so many other systems outside of it that rely on it and contribute to that circulatory system to really make this the whole system work. And having a strong analytics platform and having a foundation for AI is really building out that nervous system, the brain and the nervous system that helps articulate and say where things should happen. Yeah. And you need all the inputs. You need all the signals coming in in order to do that. And supply chain, I think, is some of the most diverse signals coming in that we can do. There's only one other area that I think has even more diverse signals maybe than supply chain, and that is our friends over in marketing. So if we go over to the next one, you know, that in many SAP transformations and projects I've worked on, the the marketing team, you know, they always want that that data about sales. They wanna know the orders, the sales, and, you know, what's really driving this customer behavior and how they can help increase their retention and increase the the overall value of the customer. How do I get more responsive to that customer? And being able to have those signals, SAP is important. It has that whole history of all the purchases that were made. But marketers use not only that SAP data that drives it, but they're using other behavioral data that's available to them. They may have Google Analytics running in the background and and feeding. How are they behaving on the in the digital platform when they're coming in? How are they reacting and responding to promotions that have sent out and what they've clicked on react to? And there's so many pieces of of additional information that flows in that has to come together to form their models that this ability to go and say, here is my priority customers or here's how I influence, And being able to tie these data applications back to their digital front end and back to you put in the hands of sales teams to be more proactive about, hey. You made this big purchase over here. Can I suggest this might be of additional value to you? This might be of additional importance to you. Right? As and I had conversation with a with a, you know, supplier of of of, know, you beverages out there. And, you know, they use AI in that to go in and see what is, you know, on the shelf and say, alright. Here's my here's all the stuff that's on your shelf. Here's where I think that you can optimize what's on your shelf. Putting that kind of power in there and say, you've already bought this in the past. These are already here. I think that you could go and and improve your margins customer and have that very customer forward kind of conversation really drives home because that is coming both from what is inside of SAP, but it's coming from what's outside. And it's creating, to your point, those data apps, those things that can help drive that decision and make that accessible. And I think that one of the really, you know, valuable pieces of this puzzle that comes with the BDC Connect and the role that BDC is going to play in here is that these data products that we're creating in Snowflake, again, aren't just sitting over in the Snowflake ecosystem. These data products that we create then are also able to be shared back. So it's not just the analyst that sees the outcome, but the operators that are in the on the ground also can see that inside of SAP because we can share these new novel data products that we created across this platform back without this massive lift in extending SAP to do things that SAP may not be, you know, from a clean core wasn't built to do, but from a business operations and up and performance can really gain traction and improve your overall outcomes. That's a really important piece why BDC Connect is so critical for an organization. So and when you go in here I mean, Freddy, what what do you think here? This in this foundation, right, what the why are we unlocking it? This is from Snowflake's point of view, where do you see this this value? I mean, Snowflake is such an important piece of this puzzle. Yeah. I mean, I think for us, it it kinda goes back to one of our, you know, core principles of being connected as a platform and being able to drive the right connections with the best in breed solutions that are out there. So we we think about this as the right partnership between SAP, which is the enterprise, you know, ERP of choice for most of our customers with a platform like ours and being able to leverage the best of both worlds to really start driving value and driving seamless, scalable AI use case generation for our customers. That that's what I am most excited about because folks are on this journey, Matt. This this just makes it that much simpler for them now. Alright. Matt, I I think you're on mute. Yep. Nope. Sorry about that. With Snowflake in here, right, and that AI foundation, Freddy, and I see this the ability to use both Cortex and both, you know, Snowflake Intelligence in here and using that from the analyst point of view and and still being able to activate and use these same kind of data products that we create in Snowflake back over in Juul and being able to access that so that this is really a democratization of that data and democratization of that AI to make this work. So that's a that foundation, and it's being able to drive both AI and the analytics platforms and be able to have better KPIs and and you work with that data across the board is so important. And then the velocity. Right? And I heard that you Gannon, you recently talked about what velocity really means. And the velocity isn't just what's my velocity at being able to do a data product data project, but it's the velocity of decision. Mhmm. How quickly can I get to an a decision and execute on that decision? And having AI, having the tools and the data products in front of people to make those decisions and have it at an an enterprise level is so important that it really ties into why this changes the game for folks. And I think that Snowflake is such an important piece of that puzzle to change that game and really get the full enterprise view of the landscape that this will impact. So with that, right, I wanna open this up to some questions that we've had come in the back and and address some of these that come a long way. So one of the questions that came in, you see, it says, SPVDC connector and Snowflake leverage the zero copy model. And let me go and just click a little bit more in here. How can you provide updates to the SAP system if no data is moved? So so this is, you know, asking about, you know, what are the the back and forth in here. And the way that the way that we see this playing out is it's about the you creating that agentic framework. The decisions that happen necessarily don't have to go about moving of data, but it's putting data in the hands. So if we let's take, for example, that agentic allocation. You're having I have orders. I have the allocation of orders, and I need to fulfill and move on to the next step for delivery and fulfillment in those orders. The step that was slowing things down, that was extending the order to cash timeline, was the amount of time it took to make the decision and then to execute on creating the documents for allocation for delivery. And using a agentic framework that then called back in and created those documents so you didn't have to manually intervene with it. So in that effect, you're you are moving data in by executing a an agentic model that puts the data back into the transactional system of SAP and performing that delivery and performing that allocation to kick off the delivery of it. And that's how this data is, quote, moving, but it's moving it because it's it's sending it along agentic signals as opposed to, you know, that big lift and shift of data that we've seen historically from an ETL. So we're not doing ETL. We're calling in using more micro activities that moves data over and back in once a decision has been made. Anything you wanna add into that, Freddy? No. I think I think you covered it well. Okay. So then we have another one in here that says if we need to extract the data from SAP HANA to Snowflake, do we need SAP Connect as a mandatory? Know, I think there's a lot of opinions on what that would be. The important piece here is I think that SAP Connect really helps do this more seamlessly, and it it really helps work with this environment easier. And that's the for any kind of ETL and any kind of I need to activate data. The goal should always be, how can I do this with the minimal amount of effort and lift? Because the value comes from how you act on the data, not the actual movement of data itself. Movement doesn't add value into there. So so that's the you there's different points of view on where you can. I think that BDC is a very strong choice on why we should. Another one that came in, Freddy, I'm gonna kind of throw this over to you is I think it really ties into that first slide, but it is we've heard AI ready before. But what's generally different here versus standard data dashboard program that, you know, becomes those multiyear efforts in here? Why is this different? Why do you see this as different, Freddy? Yeah. I I think I think that's a fair question because, you know, a lot of talk right now out there in the industry about AI and even agent tech these days, but how do you really make it real and how do you know it's real? So I think couple of things here. Because of the way we have we have designed the platform from a Snowflake perspective, it is has all the ability to take advantage of the AI and ML features that you want. Right? Let me let me go into a couple of them so you understand, and then I'll talk about what it means for our customers. But you have the ability to call all the large LLMs of your choice that you might want and run it in a secure environment in Snowflake. So we've got partnerships with everybody from Anthropic to Google. There's a whole host of library of models that you can really start using. And it's on the same data platform. So you're number one, you're not having to move things around between different parts of your tech stack to be able to do that. We also have the abilities that, Matt, you mentioned from a Snowflake Cortex perspective and Snowflake Intelligence perspective. Snowflake Intelligence, for those that might not be as aware of it, it's kind of our agentic AI interface. What it's really doing and what it's designed to do is the ability, number one, to be able to work with data and natural language and the ability to be able to leverage the right semantic model and create the right semantic model if you don't have it so that you're able to answer contextual questions from a business perspective. Now how do we know that it's real, and how do we know that it really is different than the long duration dashboarding projects in the past. We know it because several of our customers have already deployed it, and they've deployed it at scale within weeks. You know? So we have customers using us as the platform of choice for agentic agents that they have deployed to their entire organization. We have another example where Snowflake Intelligence is powering all of the commercial intelligence for a large consumer goods company. So these things are becoming real, and they're becoming real at a pretty rapid pace. And that is what really helps us get the confidence that this is something that we believe is truly game changing versus how things have been done in the past. Yeah. And I I think those those kind of pieces really, really come into play. And I think what makes this real is is well for me in in talking with customers is that pragmatic side. The the use cases that we talked about really aren't then aren't trying to go for boil the ocean use cases out there, but they're they're looking at what is pragmatic. What is the what is the right thing to do first? I think there's a question here about what what for our what's the best starter use case that proves value fast without needing perfect data everywhere? And, you know, from my point of view, what I would what I would advise any new client in this space is that, you know, we look at we look at use cases and return on investment that can go in and be easily measured because it lowers that it shortens that time frame on, like, order to cash because we're reducing the amount of manual effort to go do something, reducing the the friction in the organization to go and move a process along. Those kind of use cases build a foundation of knowledge for an organization that can then take and build upon that to go and do the same thing over an inventory. So if I'm using these kind of automations and and learning what automation works, what doesn't work in the organization, in our infrastructure, in our ecosystem, then I can go and take that into, you know, larger, bigger scale activities like, you know, the using it to order inventory and a more intelligent automated reorder process that is making sure that I have the right products that are in market, in sell in here, and not reordering your slow moving products that's coming through. And I'm making better decisions so I have a shorter lifespan on the shelf, but I'm getting that turn of my products quicker, faster. Those kind of things, I think, builds that foundation, that pragmatism as we build up as as to why this can be valuable. Yeah. I think I I mean, I I would go back to the use case governance that we talked about right right at the beginning. Alright? I think you you got to first start with use cases and and really understand, is this question worth solving? You know? If you solve this question, whatever it is, whatever priority that you outlined in your use case, Are you trying to answer something that hasn't been answered before? Or are you trying to get a better to a better answer that is going to fundamentally change the way you conduct business in that particular use case? If the answer is yes to both of those or one of those, and if you can also answer the question whether it's large enough to really move your p and l and hence worth really prioritizing, then it's the right use case to go after. To the question that was asked in terms of, like, perfect data, listen. By by its very nature, data is never going to be perfect. We are generating more and more data every minute. Right? So more and more data is gonna keep coming. It's a question of can you get to a much better answer than you have today? And if so, is it gonna move your PNL? And if so, can you get started on it? And can you keep refining over time as you get more signals and more data and get to a much better answer probably a little bit ways down the line. But at least you are moving the needle from where you are today. Alright. I think we got time for one more, Freddy, in here. I wanna pull up I I see this one because this one I I kinda with me resonates in how I in many of the conversations I have with customers. It is where do you draw the line between automation and human control? So that human in the loop, what what do you see on that, Freddy? What's your point. of view on? It's it's fascinating because I was at NRF last week, and the buzzword of the conference was agentic. Everybody is talking about agentic. But then when you start delving into it a little bit more, everybody has a slightly different definition of agentic. And then when you get into more of the details, a lot of times the question comes about like, oh, I even have the data to really do drive through AgenTic? In my opinion, I think we are probably still a ways away from pure AgenTic without any human intervention on a lot of the processes. The way to think about it is, do you have perfect information? Do you have a workflow mapping done so you understand the workflow that's gonna get impacted by whatever recommendations come out of your AI? And then do you have automation? If you have those three things in place, you can start thinking about pure automation end to end. In most cases, what we find is there are gaps in these three capabilities. Either you have imperfect information or incomplete information. If you have a workflow that's not completely understood or mapped out, or, you know, the automation is missing. My recommendation and and what I'm seeing for the most part is everybody is going down the agentic route with more of a human in the loop for the most part. For some of the more sort of internal use cases, there are agentic agents that are really driving the action and the recommendation. But for most of the commercial applications today, at least, there is a human in the lead. Yeah. I think that the the use of of AI as an assistant, as that that agent sitting next to you and helping make the decision, getting gathering this volume of information together and siphoning it into a set of recommendations. I think that's the that's that really pragmatic approach as to where this fits in because then you we see in a saw another question here about how do I trust the data. This human loop is is part of that areas that how do you build that trust? Where does that trust come from? Yeah. And that that trust only comes over time. It only comes after you've you've proven out that this is able to make the decisions, and I can trust these decisions that come along because then you you can train it to do all the relatively easy decisions. It may seem complex, but just a lot of quick easy decisions in here. But. the humanoid loop is there really to also address that that Pareto, that that 20% that's not easy, that we don't agree on, that we need to drive in. And and by using the agentic in this AI is that you're allowing your analysts a little bit more time to think through the bigger problems, the bigger impacts, the bigger risks, and they're not focused on those things that don't necessarily add as much value. They're they're valuable. They need to execute. They need to happen at volume to continue to drive the flow of business. But the human element of what is the right decision and having assistance in place to help drive that decision, I think that's where that value of AI really comes in to help drive to a better decision. And the human in there is saying, alright. This is what looks like the right decision based upon what our goals are as an organization. Absolutely. So. You're you're absolutely right. I think the the speed of decision making and the odds of success are so vastly improved by using these agentic capabilities along with a human in the loop to begin with. And then over time, as your models get trained and you have more complete information, it becomes much more easier to transition to a completely automated. Right. So with that, I'm gonna bring our our webinar to a close. And, Freddy, thank you very much for joining us today and and being part of this conversation with me. As always, it's it's enlightening anytime I get to have a conversation with you and and go over these bigger topics about where we're heading with technology with this. And for those that are on here, you please you reach out to me on LinkedIn. You can find me there. And if there's any questions along the way, find us over at dakota.com, and you'd see where we can help out. Freddy, thank you so much. Thank you so much. Appreciate the partnership as always, Matt. It was a pleasure. And thank you all for attending. And do reach out if there's anything we can help with. Perfect. Have a good day, everybody. Thank you.