Name: Transforming Treasury with AI: Coupa & Core Process
Uploaded: 2026-05-19T14:16:10.541Z
Duration: 56 min 38 s
Description: Transforming Treasury with AI: Coupa & Core Process

Transcript for "Transforming Treasury with AI: Coupa & Core Process": Hello, and welcome, everybody to our today's webinar around transforming treasury with AI, hosted by Coupa and Core Process. We still have a lot of people joining the session, so we may give them a few seconds, before we directly go into the topics. But let me start already in introducing, today's speakers. So my name is Adrien Dumont. I'm director product management, for the treasury solutions at Coupa. We will have today as well Victor Busch and Lars Beckman from core process. Then we have Hakan Backlund from Goonewho, presenting their case study. And we have a prerecorded interview with Sarisha, from Narcos Marine, as well a customer of core processing Coupa. And we're happy to share the insights of their usage regarding Coupa Treasury. So let's start with the, agenda and have a quick look. So I would like to give you an introduction of Coupa Treasury, high level overview of what it's capable to do. Then we will have, Lars, presenting around core process, how they help you, and then we directly jump into the main title story of the, webinar for today together with Gunnebo, and how they use AI cash forecasting, how they use the results in the system. That's then demonstrated by Victor. So we'll have as well a nice interview on that. This is followed by the interview for from Sarisha with regarding Lacoste Marine and how they use our, integrations to use them in their BI reporting. So this is more like, what reporting capabilities, the system has, but as well how to use it, outside of the application. And I would like to give you a sneak peek preview on, what we are working on as well, as a next step, in our releases of corporate treasury. At the end, we will have time for question and answers. So during the session, please just type your questions on the top right. You have a q and a section. Just type your question there, and we try to answer your question, towards the end of this of the session. Great. So that's that. Let's jump into the topic. Maybe just high level overview. Coupa is trusted by over 3,200 companies globally to transform their daily operations. What makes us unique is our architecture. We serve as a single unified AI platform that breaks down the silos between finance, procurement, and supply chain. By entering all of this with our comprehensive treasury solutions, we ensure your operational efficiency is always backed by total financial visibility and security. So we are excited to show you how we can drive the same value for your team today. Let's have a quick overview of what Coupa Treasury is and what it supports. First of all is the bank connectivity. So we can connect to banks via SWIFT, House to Host, and Ebix to grab your account statements with your date information, to provide you real time in a visibility into those transactions. For sure, you can use the same connectivity to initiate payments. The heart of the system is with our, cash and liquidity management tool. So here you get all your insights from all your inter entities and all your bank accounts on what's your available cash and how the liquidity changes or evolves. This can as well be connected to ERP systems to drive forecasting information as well from outside the application. Then, for sure, cash forecasting is one of the, important tools to know where your cash is changing or to which direction it goes. And then we have a module for debt and investment management, risk and exposure management, and the full fledged intercompany netting module, which is integrated into the solution. This all this is well enhanced with our in house banking functionality. So in house banking is part of all the modules and application solutions we provide within the treasury module. You can start with using cash pooling, just tracking the transactions on your, intercompany accounts, But as well, you can take it further to payments on behalf of or intercompany trading. All of this is supported as well, so you can establish a full fledged in house bank within the application. As always, you need reporting on that and analytics, which is well provided by the system. In general, the system is cloud based, so you can have access from everywhere. There's no user license fee, so bring all your users. That's, helping to drive efficiency across your organization. All of them have their insights and can drive the, accuracy of the data in the system. We have, an open system, which means, like, you can access the data via APIs or CSV integrations. But for sure, we're working on machine learning and AI, to drive further all the efficiency that we are going to see right now on the market, regarding artificial intel intelligence and where it can help. With that said, that's just an overview. I would like to hand over to Lars to present a little bit of our Core Process. Thank you very much, Adrien, for, introduction. So let me introduce our company, Core Process. So we are a strategic partner to Coupa, for soon fourteen years, and we are a full service integrator for Coupa Treasury. We have also, contracts over as a reseller partner for other Coupa services. So we have a long time together with Coupa as a exclusive partner. We are based in Sweden and Germany, supporting our network of Coupa clients across Europe. So what do we mean by a full service integrator? We, leverage our competences together with Coupa the Coupa organization. So what we do is that we bring in the Coupa technical architects when they are needed for services and together with our application consultants in the same product project. This means that Core Process can service you as one partner, bringing in the necessary services for your everyday work. We are also a service focused organization, and that is our core strategy. We service you on a hour per hour basis and use simple statement of works together with our skilled consultants who are specialized in your daily needs. We have also, aside from our implementations, which we do regularly, many fixed application contracts for the aftermarket. And including in the aftermarket where we have a large number of clients, we service you with new projects, which you will find on the references today. I'm very happy when I see the participants in the list, for the webinar today. So thank you very much for your trust, to participate. And when with the customer cases we have, we can show you some of our capabilities as a Coupa partner when we add on services to the application by using, our technology and our competence. So I now invite Victor on stage to take us further and present the cases. But, before I hand over, I want to thank you all for listening in on this webinar. And we, Core Process, look forward to support you on your everyday application work in Coupa Treasury and by keep keeping you safe. Thank you. Please go ahead, Victor. Thank you, Lars. Hello, everyone. Thank you. Good. So now now let's bring out our first interview. So we will be talking with Hakan Backlund, which is the group treasurer at a b. So for you who do not know or are familiar with a b, it's a Sweden based provider for security solutions. It has operations in over 60 subsidiaries across 25 countries and approximately around 3,400 employees. So the company is owned by the private firm, Alto, private equity firm. And these factors being across 60 subsidiaries, 25 countries, a lot of different currencies, they bring together a very complex, treasury environment. But, fortunately, today, we have Hakan Baklun with us today, and he brings over twenty years of experience in senior finance and treasury roles. So in this session, he will share how AI can transform cash flow forecasting from a more manual time consuming process into something more faster, reliable, and maybe a little bit more strategic. So, Hakan, please join us on stage. Hello. Thank you for having me. Hi, Hakan. How. are you? Not too bad. How are you? I'm very good. Thank you. Good. So let's let's start with the interview. So why is cash flow forecasting so critical for goodwill, Abi? For us, cash flow forecasting is absolutely critical. We are a private equity owned, so liquidity and debt managements are very front and center. We need to know where we stand on a day to day basis. Reliability is, of course, the key. The forecast, if the forecast isn't trustworthy, it's hard to act on it, especially when you are making decision quickly. It also supports bigger decision, investment, acquisitions, and because you can plan and then you can plan the funding and timing with much more confidence. And. since, we operate through many subsidiaries across different countries, we need one consolidated view, not just a local snapshot. So finally, in some countries, there are tighter liquidity and currency restrictions. So being able to predict and plan the cash movements, becomes even more important. Okay. So before introducing AI, what were the main challenges you The biggest challenges we was challenges was reliability. faced? We had many subsidiaries sending input and not even when everyone put a lot of effort into it, the quality still varied. It was very manual and time consuming, and in the end was really hard to fully trust the result. So we spend a lot of time double checking. Okay. K. So then what made you then start looking at AI as a solution? We saw an opportunity ought to automate a lot of the work and improve the quality at the same time. Instead of relying on manual inputs everywhere, we wanted to use patents in our own history payments behavior to create a more consistent a constant baseline. There were that's where we saw the real potential, more reliable forecast with much less effort into it. K. So now let's go in and talk a little about how this actually work in practice. So but not going too deep. So still at the high level. So the model is based on historical data. But from your perspective, what is important to get right, in order for this to work? For us, the most important thing was having a solid historical data and making sure it was clean and constant. The better the data, that's the better forecast. And one thing we worked a lot with was structuring the data, categorizing the into payments into different categories. And yeah. So from your perspective, how important was this step? It was very important. When you can group the cash flows, customer collection, supplier payments, salaries, taxes, becomes easier to build something you can trust. The structure is really the foundation, Yeah. I would say. Exactly. So once the structure is in place, then the model can start learning and even then automatically categorize new transactions over time. And yeah, so how important was your treasury system and then enabling this? It's it's essential. You need a place where you where your transaction his your historical transaction and cash flow data collected and governed properly. Without that foundation, the the forecasting doesn't work. K. So let's move on to the results and the impact of this. So what kind of results did you see when you first, tested the AI, cash forecasting model? In our earlier test, we saw we said that anything about 75% accuracy would be acceptable. What we saw was typically between 8096% for selected entities, so it was much better than expected. And compared to your previous way it's working, what has now changed with the AI cash forecast by now using the AI cash forecast? The big change is speed and confidence. We get the forecast much faster, and they are much more consistent. At the same time, we are reducing a lot of manual work and the subs for the subsidiaries. So it's really both an efficient gainer and a quality improvement. K. And how does this impact decision at management level? Gives the management much more confidence into the numbers. When you can trust the forecast, you can make decision faster and with with less risk. And that's the real this where the real business value shows up. And, there's a lot of discussion about AI replacing people. How do you see that in treasury? I don't see AI replacing treasury. What it does is improving the repetitive manual work. But just but you still need people for the analysis and the business understanding. We need to interpret the numbers and adjust things for, for for the model, that the model can't know, like, one of some sudden changes. So so you would say that AI, it's it's not really a a magic wand. Exactly. I think, of it as having a very fast and capable statisticians. It can produce strong forecast, but it doesn't understand the business context. That's why human oversight is still essential. And how do you then then handle this unexpected events or, outliers, in your cash forecast? There will always be external events that affects cash flow. The difference now is that we can start from a much more reliable baseline, and and then we can adjust rather than building everything, manual from scratch. So yeah. Then I think we should also ask a question about the security. So a question of security. This is very sensitive, financial data. How do you ensure that this is handled safely? That was the one of the key requirements from day one. We enrolled IT early, and we made sure that the solution sits inside our secured environment with strict access control. So both the data and the solution are protected and not available externally. K. And, finally, what advice would you give to companies considering AI for cash flow forecasting? Start with your data. Make sure it's structured, consistent, and accessible. Then test the test in a pilot before you scale, and keep in mind that AI support decision making, not replacing it. Okay. And could you give the audience maybe a final sound bite? So if you summarize this into one sentence, what would be the real value of an AI cash forecast? For me, the value is simple. Faster and more reliable forecast and more for more time for analysis and better decision. K. Thank you, Hakan. Thank you very much for participating in this interview. Thanks so much for for I'm so happy to that you have been here. Thank you. Yep. Thank you for having me. Thank you. Bye bye. Good. We will start with some poll question here, and, you can answer in the in the poll as you see on the far right. So how is your cash for for forecasting process managed today? So please type in an answer here to the far right and, see how let's see if we can see how the audience is handling their cash forecast. Oh, there's a lot of fully manual work I can see here. Fully manual. Yeah. That's the leading that's the leading part. Yeah. Still. But there are some we do still have a lot of automated. Well, now, let's let's move on to we have also another poll question. So most of you, you handle it it, a manual today. But let now see what is your biggest challenge in cash flow forecasting today. They have one vote for manual work and time consuming. Yeah. It's quite even there. So it's this forecast accuracy, manual work, and data quality. That's that's the the a lot having a tough okay. So, interesting. Very interesting to see that, answers. But now we I want to show you in Coupa treasury. So showing you the actual, data and, how it looks like in Coupa with the AI cash forecast. So I will start sharing my screen. Now let's see that I'm not logged out yet. Good. Yes. So this is how it looks like in Coupa Treasury for for you have not seen the system yet. So when you log in to the system, you have your, dashboard in front of you. And I will talk about the dashboard, later. But for now, I will keep it as it is. And on the left hand side, we have the different menus within the system where we can go into the the different functionalities in the system. And on the far far right, we have our ai navi agent as well. But now we will look at the ai cash forecast. So we will start by looking into the the the values that actually creates the cash forecast. So how is the cash forecast the AI cash forecast created? Well, all the banks, they send data into Coupa. So all the account balances, with them contains, the transactional information that is in that is sent into Coupa Treasury. And in here, as you can see, this list contains a list of 13 transactions for various accounts and with different values. Each transaction contains, information from the bank such as if it's a VAT payment, if it's an invoice, and and so on. And by looking at this information, the system can use machine learning, so categorizing the data into different categories. So as you can see, the system has then categorized the categories into different categories here on the right. And what this does in turn is that AI then pulls this information and creates a cash AI cash forecast based on this information and then pulls out forecast data in into the system. So the the the AI forecasting data is then pulled then back into the system for the for the actual each forecasted flow. And this could be set up as it could be done once a month, every day, every week. Whenever you want the AI to update it, you can set those settings, inside the model. And once you have, once this is done, you let's take a look at the the balances and forecasting interface. So the AI sends in the data into the system, and then we will have a report. And in here, we can then see that we'll see an a graph. So starting from today, we have our bank balances as the consolidated group bank balances. And every day, we can see, AI doing a cash forecast here, so transactional data here on each day. So changing our consolidated group balance of the group. And down below, we can see a table. So this I've just included three days. So starting from today, tomorrow, and the next day after, and how our balances move. So this is the balance as of today, but as you can see, the balances, the group balance, it changes every day. And we can also see that it's on a currency level like this, And then we can see how our currencies move day by day, but also consolidated into our group currency. And in here, we can drill down so we can take a look at all the different accounts. So in this case, we can see that this account here, this SAK account, is negative today. So we're paying interest to the bank, which we do not want to. So here we can set up a transfer. So in Coupa treasury, as Adrien said, we can connect the banks. So we can do payments from the system. So in here, we can say, well, let's fill up this account. So we click on the account, and then we create a transfer, and we can see our balance here. And let's cover this position by transferring 500 s k k. And now we will have a new balance of 72,000. And then we select we want the payment reference. We say internal transfer. We select execution date, so let's say today. And then we select which account. And then we save this transaction, and then that will be filled up this specific account. The cash forecast is also done. As you can see, it's it's in each currency, but this can, of course, be changed. So let's say that we want to have a look at it on a entity level. How is each entity performing day by day like this? How is, each how is the cash forecast based on bank level? So we can see how is Nordea, how is SEB, how much cash do we have inside each bank day by day like this. And here is also easy to insert a chart. So let's say that I want to insert a chart, and I add my cash position. And now I have a chart based on each bank. And it's also easy to change, like, if I want to see this on a entity level, for instance, like this. It's also possible to drill down to see it on a category level. So the AI performs, the the cash forecasting on a on a category level as well. So if you're paying salary every month at the 25, let's say, that you do in Sweden, the AI will perform a a a forecast on that specific date, and then you will be able to see what are my VAT payments or salary payments and and so on. So if you drill down and have a look at this so I go click here, drill down, and then I can see this is my cash inflow. This is my cash outflows, and this is based on each category. So this is my VAT payments. This is my salary payments, and this is the payments that I have during these three days as you can see. And here it's possible for me to do changes on these values. So let's say that we hired a new employee, the salaries has in increased, well, then I can add a new value here. This also enabled that well, let's say that you do an AI cash forecast, but you also want the entities to log in this you want the users to log in to the system and double check the values. Then it's possible for users to log in here, and then they can review the forecast. And if they need to do any changes, they can do it here. There might be also, that you want to freeze the period. So let's say that the AI or the AI with the help of your user created a forecast, and you want to freeze the moment. So you have done a cash forecast now in May, but you want to freeze that specific period and make and then do a comparison later so you can compare what did I forecast versus what, what did the bank say? Then it's possible to do here in Coupa. So you go back to my, shortcut here, and I'm we run this report where I have the the month at first. And then in the a column here, I have all the actual bank transactions coming in from the bank, so on a different category level. And in the b column, I have all, the forecast, transactions. And then it makes it possible for me to freeze the specific moment, and then I can do that comparison between those two. So always enabling seeing the deviation between actual versus forecast. And, of course, there's a lot of extra functionalities such as you can drill down on these numbers and so on, but that we can save for another day. Now, I want to move on, to our next, case, which is with, our, NACOS marine case regarding a BI tool. So, regarding that, so it's the the interview with focus on the ship automation and navigation group, Nackers Marine. The group is a spin off from the Finnish industry group, Wetzila. So started last year, the company had to migrate all new data to new company structure, And this meant challenges for their bank structure and group reporting as new measurement has to be accounted for. So please, have a look at the recording with Suresh Srimana here. Okay. I'm joined by Suresh Srimana, business analyst and BI developer at NACOS Marine, who has been instrumental in building the company's Power BI environment. Suresh, you helped, connecting Coupa Treasury with Power BI and other system to create a centralized reporting layer. It's really great to have you here today. Thank you. So let's start with the background. What was the situation before you built this Power BI? So before Power BI, our reporting landscape was more scattered across multiple systems like treasury, ERP, and CRM and others. Mhmm. So it required a significant manual effort and time to consolidate everything together, And that made it, difficult to achieve a holistic and a real time view of the business performance. So what was the idea behind, then introducing Power BI to the group? So our goal was to establish a unified platform that consolidates data from, various sources into a more consistent and visual format. And, Power BI enabled us to merge treasury information from Coupa along with the ERP data and other data sources. So this gave us a comprehensive and a unified view across the organization. And not going into much technical details, but on a technical perspective, how does integration with Coupa actually work in this case? So we are using, APIs to pull the structured data from Coupa treasury and then integrate it with the, ERP data within our central, data warehouse. And since Coupa's data is already well organized, the integration process into the Power BI is just very, seamless. So then you would say that Coupa is the foundation of all the treasury data in Power BI. Yes. Absolutely. So Coupa still serves as our record source of, treasury data. And Power BI just complements it by, or rather than replacing it, it just complements it by extending its value by making the data more accessible and as well as connecting it with other broader business information systems like ERP. And what would you say was the biggest value of the setup with Power BI? Well, the sign significant benefit, is having a centralized and a well governed, reporting framework. So we have eliminated much of the manual work, and achieved greater data consistency and as well as enable the users to access the insights more quickly. So now our time goes into, analysis rather than the report preparation itself. So about the governance structure, could you tell us a little bit more about that? Why why is that important? Yeah. So this centralized governance, approach, guarantees consistency and as well as scalability across all reporting. So our dedicated BI board, evaluates and approves new requests, which helps us to maintain the data oversight and prevent any kind of data fragmentation as the organization is expanding. And, also, it helps us to maintain a a single source of truth for all Power BI reports, across the, business to steer the business. So we're coming to an end of this interview. But, if you could give us one piece of advice to companies starting a similar journey, what would that be? So I would say, maybe first prioritize the solid data architecture and structure and as well as governance from day one. And when that foundation is in place, building a scalable and a high value b BI solution, it just becomes more significantly easy and as well as more straightforward. And if you would summarize this into one sentence, what's the real value? I think it focuses on just integrating data from multiple different systems, into a unified and a real time view to just enable making smarter business decisions. Okay. Thank you very much, Theresa. Thank you so much for participating here. Thank you. My pleasure. Thank you. Good. So that was our interview with, Sarisha at Nackers Marine. Now we will have another poll. So let's see what the audience, says about what is your biggest challenge today in financial and treasury reporting. Are you, doing manual tedious work or are your data spread across different systems or what what is your how how are you working today? Yeah. It's a lot of manual work. I can see I can see in the in the votes. Yeah. And the the data is also spread across multiple systems. So it's it's a lot of, manual work still and, data spread across multiple end systems. I think those are the main drivers here. Very interesting. Okay. So now for you who have not seen a bi tool before, I was thinking of just sharing a a quick slide how it might look like. So what, Sarisha was actually talking about. So, I mean, this is an example how it can look like once you create a report in in the BI tool. So for instance, on the left hand side, you have your calendar, which you can click on any date, and then it will see the data for that specific date. And then, as you can see in the in the middle of the screen, there are the different values. So you can see that there's a lot of data coming in here for the treasury application such as the liquidity, but you can also see, for instance, that the guarantees, which it could be that you maybe have another guarantee system to handle guarantees, or they can also be in your treasury in Coupa. So other or it's it's then still can be managed into the BI tool to pull specific reports such as guarantee utilization versus your credit lines and and and so on. While things such as matching the treasury system versus your ERP balance, as you can see in the the in the far right corner, it's something that just, of course, needs to pull information both from ERP and the treasury management system. So just to give you a slight taste how how it can look like, but, of course, it's still what does your management want to look like? What is important for you to look to look at. These are all, things that you can create on your own in a in a unified, BI tool. So I will stop sharing that, and I will log in to, Coupa once again to show you a bit about how the dashboard, functionality and reporting functionality can look like here. And, what we can see now, is the dashboard. So as, as you can see in here, I will zoom in a bit, we have a a in the cash AI cash forecast, which I showed you earlier. So starting balance as of today and how the balances moves. On the right screen, we have a reconciliation functionality in Coupa. So Coupa can look at your account statement, and then they can tell you, are those statements up to date? So can we actually, trust the balance inside the system? So this is what this tells us. But there are also other, customized, tables you can add according to your needs. So for my case, I added, what are my priority payments in the next couple of seven days? What are my credit lines? So is there any headroom left for my guarantees or overdrafts? Yeah. You name it. But also, what are my, approved payments as of today? What will leave the bank today? But, also, according to how much cash do I have in each bank? So can I actually pay this amount to the right? Do I have enough funds in the bank to pay this? So there are different values you can different charts you can add here to your customized dashboard. And, of course, it's all that you can have different dashboards. So maybe you're not interested in into the cash or maybe you're just a payment user coming into the system knowing, okay, what are my payments, that I need to approve? Then you can create the dashboard just adapted for payments. So let's say that you have today's payments, which, are going out, but also which payments has not yet been approved. So in this case, there are 30 payments would not yet been approved, but also the the balances. So can I actually pay these payments and and so on? So these are various, charts and and data you can pull from the treasury system to to give you a little bit of, IDs, really. And as said with the Power BI and as as NACOS users, they use a lot of APIs to pull the data as they want to report in one unified, system. Then in that case, they go, they can use the API documentation. So here we see a lot of of a long list of the possibilities within Coupa treasury to to pull and and both insert information into Coupa treasury. So it's a very low extensive list. Good. Now we'll hand over the word, to Adrien, which will talk about the Navi agent, the AI agent in Coupa Software Treasury. Exactly. I thank you very much, Victor. Impressive insights you have granted already. I thought, okay, this is a reporting session here or a case. Let me as well bring an idea of how this could as well work with AI. So just to, be honest, this is a a preview of, something which will come up in the next release. That's the current plan. But just to give you already some insights into that. So once you have, Navi enabled, I think there were questions around that as well for those who are already Coupa customers. Please reach out to, the support team. They are happy to help you in enabling Navi if you do not see it. But, I just wanted to share one, idea of, for example, just asking Navi, to summarize, my today's account balances, per currency and country. And I just send that to Navi, and, let's let's wait while Navi is analyzing my data in the background. Right? So it finds all, already that within our agent portfolio, there is a dedicated agent for cash management and, tries to retrieve the correct data from the application. You will always get a status update on what Navi is currently doing, which is very great. So you see it's still working and, not locked out or something. Yeah? And this is just like natural language that you put in, to ask a question, to get something out of the application. And now what we see is that I get a report of my account balances for May 19, so for today. And I just can click on this, and I get the documents around, what Navi has figured out in my application. Right? It really gives me an overview of per currency, what's my balance. It even shows me the accounts. Then spreads it across the the the countries that have this currency. And I can really scroll down the list, which is just one way of, getting into that. And I said, this is a first start of this. Right? It's just a preview, but it will come, in one of the next releases, and we are happy to get as well your insights and your feedback, if this is helpful and what we would like to see next. Something that even surprised me is maybe towards the end of this report that Navi brought in some key observations with what's my largest positions or where I should maybe act on, like, my negative balances. Right? So this was not explicitly asked, and there was no real guidance on that, but still this feedback came as well. So I'm really much impressed on what AI can as well do for the reporting capabilities. And you just see that there's, for example, other options that Navi is already providing me, like, would you like to compare it with previous period? We can as well click on that, and then Navi is running again and will run reports on how your balance is changing, and give you insights. For sure, all of this is as well doable directly within the application and you are maybe a user of the application, a power user, so you're quick there. But just imagine all the local or the CFO or whoever is not logging in that often can just talk in natural language with the application and get the insights they would like to get. Yeah. So this is still running, but it will put out a similar document and just compare it. Good. Perfect. Thank you, Adrien. Thank you very much. Good. So I think we have some time left for some, question and answers, or I hope answers at least. And see. I can see that already coming in some some questions. There are quite some questions exactly. Maybe, if Hakan, if you're happy to join us on stage and, to answer some of them because some are addressed directly to the first AI use case. And maybe yes, Victor or Hakan, you just can answer them and I will go through them. Cool. So one of the first questions was, what is the precision of the forecasting related to the timeline? Like, is it, like, 80% or 90% accurate, or have you measured that? You you can you can measure a lot. I would say that, yes, we follow-up on on, as as I said previously, it's, we are following up on per entity of the outcome. And, as as we say that the 80 to 96%, that's sort of where we are from the from the worst to the best. And, of course, it's also a size of the entity, smaller, less eight day or intensive companies, easier than a heavy, and so on. Okay. Yeah. Yeah. I think that's always a a big question around, okay, how helpful. is if it's not accurate. Right? Yeah. Then there. was another question, which is very interesting, I think, regarding the time horizon. So how far into the future are you planning with AI and on. on what granularity are you planning, like, per week, just once per week, or like on a daily basis? Yep. We are doing it short term liquidity analysis. That's our main point for it, and we are doing it on a weekly basis. So we are filtering it in in into weekly back buckets. Okay. If and that, when you say short term, temp depends on company to company, but we are short term liquidity analysis. Great. Thanks for that. Yeah. Victor, jump in if I k. Yeah. But that's that's that's that's a good answer. I think it's yeah. the the time period is was it two months you're. looking at mostly? Yeah. Yeah. Yep. Around? Yep. Great. On a rolling. basis. Sure. Yeah. And then we get the weekly output. Got it. Yep. Great. Yeah. Then there was a question regarding, the data source. So is it like actual outgoing incoming bank payments, or is it ERP data or Yeah. No. APAR? We are taking the tip no. We are taking short term APAR, numbers from, from Coupa, the TMS system. So classifying payments, I said before, it's, it's essential into into the into the analysis and get the the the data right. Because if you have a lot of of noise into the system, of course, you get noise out. So classify your payments, structure payments, and classify your your your payments, then it's a sort of cleaner output. Of course, there is always be noise into the system, but the more you can work on it, the better it is. Sure. But to be very concrete, so you really use the cash flows, the actual cash flows out of Coupa Treasury, from the history, Yep. and that you feed that into the. AI. Right? And. whatever, that. think, is in there. Yeah. yeah, I think you need to be the sort of historical data. Here, we take it depends on the on the time horizon you should take back. I think you can do it, twelve month rolling data. I think that's the the best. Correct correct me here, Victor, if I'm wrong. But then you get if you have twelve months or or fourteen months, then you get sort of all the the last twelve months into into your, forecast. If you get more more data from from sort of more historical data, you you add noise and you add changes into the system. So I think, twelve or fourteen months on a rolling basis. Again, Victor, correct me if. I'm, if I'm wrong. actually, it's it's actually a lot better with more data. So the more. data you got, the better, decisionalities, as well aligned, so you can. see the seasonal effects, to it. So, I mean, I would recommend at least three year of data, of historical data before doing. AI cash forecasting to get really good values. Of course, you can do twelve months, but it won't, I mean, for seasonal effect, it won't get that good. Yep. But, yeah. But but on a rolling basis. Yes. Sure. On, a. Sure. rolling basis. Yeah. Yeah. Of course. Exactly. Yeah. Yeah. Yep. Exactly. On a rolling basis then for the. to increase the the weights, so to say. Yes. Then there was one other question, which is as well very interesting, I think is how long does it take to implement something like an AI forecasting? Right? So for sure you had the data already in in treasury, but, then from there to really getting something actually useful back, how long was this time horizon? Well, it's I I would say it was I was surprisingly quite short, to be honest. I I I saw a longer sort of, longer, sort of project, plan in front of me, but it was quite fast. I I'm not going to say month. I I say month. It's not, not long. So if you have, again, if you have your data in in one system, if you have it, in in a sort of orderly fashion, it's it's easier. Then it's, of course, how you want to structure your your forecast and what you need it for. So, of course, our I say it is quite simple because we need it for for for liquidity and liquidity planning. But, you can you can add more into it, and then it gets more complex and all that. But we are quite easy straight straight through firm, working with projects, and then it's sort of quite easy, I would say. Again, Victor, you you helped out with this, Great. of course. Again, Yes. Yep. But might think another, view. very essential that you had a a lot of data already, and this was very structured. So that was I think that was the key, for a fast paced project in this, this time. Great. Yes. That have been more or less the questions to you, Hakan, and, to the AI. I. saw I saw one here. It's revolving the governance. And, to say that here, you have to take into account to to yes. We are we are sort of in a in a close environment. Then, But again, we are private equity on firm and we can we we don't have to think about, sort of, what we can leave out and what we what we can't, as we don't have or listed anywhere. So that the it's, of course, regarding governance and and implementation in the AI. It's it's good. And also, we are a smaller firm. It's easier to to, to have the decision, in in quite, quite fast. So and I understand if you are a larger company, it the governance will, will take a bit longer. Absolutely. I think, that's an. important part. Of course. Yep. And, I mean, you can still do changes to the four calls. So you can have the AI doing the layer, the groundwork for you, Yep. and then, Yep. I mean, you as a user log in to the system. And then okay. Well, Yep. yeah. The salary said the taxes are are correct, but I think the invoices will be at this level this month, Oh, for instance. yep. Exactly. Good. Great. So maybe, Victor, we have now some interested, attendees on this AI forecasting for sure. They would like to see it a little bit more detail or get more information. Shall I just reach out to you? Sure. Yeah? Absolutely, Okay. please. Yes. If you have any questions, please reach out to me. So Coupa's happy to help you, and and giving you further insights in a in a one on one demo. Great. Good. Then let's jump. I think there was one BI Power BI question. So, Mhmm. the question was regarding does Coupa offer Power BI templates? There, I need to say no, not directly. But we have partners like Core Process. So happy to help you with the experience they have, but as well our implementation team. So, just reach out to to to Coupa our core process. We're happy to help you there, and share our experience with it. Great. And then there was one question regarding, for Nucleus Marine. So how long, did the implementation take for, the BI reporting solution you have developed here? Yeah. Do you have that, Victor? Yeah. It it depends, of course. I mean, in their case, I mean, it was quite I mean, it to create a dashboard is quite quick, I think. But, I mean, if you want 10 reports, of course, it it will take time, to create that. But I mean, we will, yeah, we will take a look at the data and then we will give a timeline. But so I think that's I mean, looking at maybe weeks of maybe week until it's it's finished, so around to give a. guesstimate to create a a more like a a dashboard or something which is complete, which is then, connected to to various systems. Yeah. Great. I think these have been the dedicated questions. There is still a lot regarding Navi now coming in. We had maybe a good preview of what's coming next. So let me just talk a little bit about Navi itself. So Navi is coming to the Treasury application with the r 45 release, which is currently rolled out to all the customers. But your core environment needs to be r 45 as well. I think that might take time. If you do not see Navi, there are multiple reasons, so you need to have the roles, permission, entitlements. So please check the documentation, and then you get it activated. And what is, Navi able to do is, with the r 45 release, you will get access to the, Compass documentation that's as well a question here so that your users do not need to, yeah, go through the help pages, but they can just ask a natural language how to use something within the application and then they get, the feedback directly from Navi. As I said, that's the r 45, the current released version. And then, we shared that as well in our, r 45 webinar, regarding the roadmap of what Navi will be able to do. So with the next release r 46, we would like to ship what I just have demonstrated regarding, some balances inside. So at least to get, a read access to all the, cash management, data objects for treasury so that you can report on those. And then we will enhance that to all the objects that we have as well integrations for. Right? So that you have as well even insights into money markets, financial instruments, into netting. Wherever you would like to get access to, you can as well use that kind of data conversational to get some reports out of the application. That's just part one. The next part is then to modify the data. But stay tuned. That's future, maybe 27. Good. Yeah. I think. that's it. Most questions answered or even all. I think that was a great session. A lot of insights. Thank you very much for that, Victor, Thank, you. as well to, Thank you. Sarishta. Thank you. And yeah. So to the attendees, please reach out to us. We are happy to give you further insights. I think there is even a nice slide where you have some QR codes if you want to find us on LinkedIn. But as well, just use our emails or why the general connectivity message. Good. Thank you very much. thank you very much for participating. Thank you. Thank you. Bye bye.