Skip to content

Euroports - Direct/Indirect Bridging Prototype - 2026-04-15

Metadata

  • Date: 2026-04-15
  • Company: Euroports
  • External Participants: Matthias Depoorter (Treasury Manager)
  • Palm Participants: Emma Sjöström, Gurjit Pannu
  • Type: Lead Call
  • Domain Areas: Bridging, Cash Forecasting, Categorization, Investments-Debt (factoring)
  • Recording: https://tldv.io/app/meetings/69df9a238343c3001364d74a/

Summary

Context

Follow-up prototype review with Matthias (Group Treasury Manager) on the direct/indirect forecast bridging concept Palm is building. Emma walked through an early v0 prototype; Matthias shared his current Excel-based bridge and the board presentation format he has to produce each quarter. Gurjit framed the session as "discovery and learn" — not UI validation but confirming which outcomes and information matter. Contract/VDD acquisition process still in progress on Euroports side.

Key Discussion Points

  • Matthias walked through his current bridge (tab "Cash flow analysis CA vs BU") — the canonical shape of what "bridging" means to him
  • EBITDA / EBITDA adjustments come from budget; working capital rows must come from cash (transactional) data
  • Main focus is always working capital deviation — what drives the variance between direct and indirect
  • Current process: variance numbers derived from Cash Analytics (manual categorization), explanations obtained through "lengthy interview" with local entities
  • Board presentation is the output artifact — consolidated quarterly view with delta vs budget + written commentary. Needs to come directly out of Palm to avoid manual PowerPoint work
  • Need to drill from consolidated → entity → transaction level to explain variances
  • Local entities need their own view of their numbers vs budget (educational/collaboration surface), with access rights segregation
  • Sub-consolidation grouping required (Italy = 2 entities, Belgium = 6 entities, Holdings = 5-6 entities)
  • Forecast bias from local entities is a real problem (AP people overestimate payments) → target overlays needed
  • Factoring is a major working capital driver — ~50% of AR globally factored via ABN; only entity-level factoring % visible today. Wants customer-level drill-down on factoring data
  • Budget is annual (made once), forecasts done 1-2 times per year — direct cash forecasting must bridge the time gap
  • AR/AP granular data not available for all entities (immature entities) — Matthias investigating with IT/accounting to get SFTP feed or ABN API

Pain Points

  • Manual bucketing in Cash Analytics with frequent miscategorization (e.g., finance payments booked as operational)
  • Cash Analytics categories too coarse — "cash flow from operations other" is a huge bucket
  • No drill-down from variance number → driving transactions
  • Bridge must be manually rebuilt in Excel each forecast cycle; commentary captured in free-text box
  • Explaining working capital variance is the hardest part — "that's the complexity of bridging a booking-based vs purely cash world"
  • Local entity personal bias in forecasts (AP overestimates payments, conservative entities underestimate collections)
  • Factoring payouts arrive as lump sums — no visibility into which customers drove the factoring %
  • Two ERPs to reconcile (terminals on Oracle, freight forwarding on InVision)

Feature Requests & Needs

  • Attention-grabbing top-of-page variance flags pointing to where direct/indirect are drifting apart (especially working capital)
  • Category mapping UI: each indirect (budget) bucket → multiple direct (Palm category) equivalents, showing "how much is coming from where"
  • Drill-down on variances: supplier payments −3.3M → structural/stable/unfavorable breakdown → largest transactions influencing it
  • Entity-level views with group/sub-consolidation (Italy group, Belgium group, Holdings group)
  • Commentary capture in Palm — replace the free-text box in Excel; Matthias + Evelyn should be able to add comments per variance
  • Forecast accuracy KPIs per entity — "scoreboard" for CFO showing which entities forecast accurately vs poorly
  • Target overlays for entities that systematically under/overestimate
  • Version locking of forecasts so next forecast cycle can compare against the previously-communicated version
  • AI-assisted commentary: select from auto-generated explanations ("major influences: X, Y, Z"), then push selected commentary into the report
  • Report-ready output matching the board presentation format (actual, last estimate, previous estimate, prior year, budget columns by quarter)
  • Factoring drill-down: from entity-level factoring % → customer-level contribution

Jobs & Desired Outcomes

Job: Bridge direct cash forecasts to indirect (budget) forecasts and explain the variances

Desired Outcomes: - Minimize the manual effort to reconcile direct cash forecast totals against indirect budget line items - Reduce the time to produce the quarterly board-ready bridge + commentary - Increase the granularity of working capital variance explanations (from entity-level to transaction-level) - Minimize the dependency on interviewing local entities to understand variance drivers - Increase the reliability of category mapping between bank-statement categories and budget line items - Reduce the volume of miscategorized transactions polluting variance analysis

Job: Drive forecast accuracy and accountability across local entities

Desired Outcomes: - Increase forecast accuracy at the local-entity level over time - Minimize the effect of personal/role bias on submitted forecasts (e.g., AP overestimation) - Increase local-entity ownership of their own direct vs budget variance - Reduce the need for group treasury to manually overlay targets onto local forecasts

Job: Explain factoring-driven working capital movements

Desired Outcomes: - Increase visibility into which customers drive changes in factoring percentage - Reduce dependency on PDF extracts from the factor (ABN) for customer-level data - Minimize the effort to reconcile factor payouts to individual invoices/customers

Domain Insights

  • Board presentation structure: columns for Actual, Last Estimate (latest forecast), Previous Estimate (prior forecast), Prior Year, Budget — always by quarter
  • MPL vs Non-MPL split: terminals business (Non-MPL) gets detailed variance breakdown; freight forwarding (MPL) is one line
  • Bridge mechanics: EBITDA + EBITDA adjustments from budget directly; revenue (row 6) and costs (row 10) feed EBITDA; working capital split into trade receivables, trade payables, intercompany (should net to zero consolidated), and "other" (accruals)
  • Factoring mechanics: entities upload invoices daily via "The Basics" tool → ABN accepts/rejects based on credit insurer (Atradius) limits and age (>90 days penalized). Daily payout. Booking: factoring advances posted as negative against trade receivables; 5% SPV released when customer pays. Goal: net-to-zero across trade receivables / factoring advances / SPV
  • Two ERPs: Oracle (terminals) + InVision (freight forwarding)
  • Working capital variance is cash-driven — the defining challenge is explaining changes without accounting accruals noise
  • Budget granularity is the ceiling — direct forecast can be more granular, but bridge must roll up to budget line items

Action Items

  • [ ] Matthias: meeting with accounting this Friday + IT this week to explore AR/AP balance data feed and sub-category P&L splits
  • [ ] Matthias: pursue ABN factoring platform access (token went to Luxembourg in error, needs resend)
  • [ ] Matthias: send feedback to Emma/Gurjit after Friday meetings
  • [ ] Palm: keep budget bucket → Palm category mapping in mind during tomorrow's categorization workshop with Giannis
  • [ ] Palm: follow up in ~1-1.5 weeks with progress for testing
  • [ ] Consider testing AI chatbot commentary on bridge concept by loading category mapping into Pulse (temporary session-level test)

Notable Quotes

"I want this file in Palm." — Matthias (earlier quote, referenced by Emma as the reframe for the session)

"We should be able to split that up saying, it's not really Revenue in general. It's storage revenue or it's thermal handling collections... Much more granular. Now I cannot go into that detail based just on what I'm doing in Excel. So what I'm hoping for is to get that granular data." — Matthias

"For me, what I need to answer is always: what's the working capital deviation? What's driving that? Is it collections speeding up? Is it collections slowing down? Are we repaying our suppliers too much?" — Matthias

"What the board is looking for is actually explanations that are cash driven. And this is the difficulty. That's the complexity of bridging a world which is booking based versus a purely cash pay." — Matthias

"What you're seeing here is a consolidated view across the whole organization. But what would be helpful is if you go from consolidated to entity level and then ultimately transaction level." — Gurjit (summarizing Matthias' need)

"Depending on who does a forecast, their personal bias influences that. If somebody from AP is doing it, obviously they overestimate payments because they want to make those payments." — Matthias

"It's much more than purely a cash flow forecasting tool. It's also an educational tool, at least for the local entities." — Matthias

"This is more than enough validation for us to keep going in this direction." — Emma

Full Transcript

Them: Hello. Everyone. Hey.
Me: Hey!
Them: How you doing? Good. Good, good. Busy, but good. I guess everybody's busy. Always good. I think being busy versus bored is a lot better for sure. Yeah. Let me get in this piece here. Emma, can you hear us?
Me: Yes, I can hear you trying to get the video, the TLD week going, but.
Them: Okay, I gotta. I got it in there. I got it in there. Very good. How was the. How's the skiing, Matthias? Good. It was a nice weather in Italy, so. Nice. In the morning also the snow was good. In the afternoon. It was less good, but then we could do other stuff. So I didn't realize it was still snowing at, like, mid April in Italy. But I don't know much about. Like, I think a week ago or two weeks ago, like. A meter and a half of snow. Or nothing. Because Easter holidays is more difficult for skiing, but it's easier for the kids. So. Oh, there you go.
Me: Yeah. Fair enough. We're in Italy like the Alps did you go?
Them: It's called Grana. So it's like the domites. I think.
Me: Nice.
Them: Yeah, I have to. You have. You have to pass through the. The mom tunnel. Which was. Yeah. To go through. Yeah, but it was still doable. I mean, I left at. Three, 3 a.m. in the morning. So I was there by 11. So it was okay. If you get there after. After 12, it's more difficult. But when I had to. When I went back, I didn't have any issues. Because I. I traveled on. On Sunday, so most people leave for skiing and return on Saturday. So it's. It's good to avoid the peak moments.
Me: Fair? I've only gone through that tunnel by car once. And I think we were quite lucky too. We only had to wait an hour and a half to enter it.
Them: Okay. That was. Yeah. For me, it was only half an hour. I can't really complain then.
Me: They were like oh you know people wait for three hours. We were like oh wow. But that's good super cool. The apps are really nice.
Them: All right. Nice. Thanks for. Thanks for making the call. I won't jump into the red line and contract questions. I'll let you keep that with the. With George. Yeah, I've already said multiple. Multiple reminders, but we're in a. We're in a, like a VDD process. So potential Acquisition. So legal is. Is more focused on that. And then one. One. One of my legal colleagues, she actually had a car accident, so it's only one. Of two. So I don't want to ask the guy too much. By end of end of this week, so I'm. I'm hopeful. Okay, cool. Yeah. Yeah. George shared the same. And I told George I won't. I won't push for. For. For answers there. I know that YouTube are both on top of that, but I think today, totally different topic. I'm gonna let. I'm gonna run this, but overall, like, look, we're kicking off the work on the indirect and direct forecasting bridging. And so I'll. I'll preface this with it still pretty early. We're very much in learning and sponge. Being sponges mode. And a lot of these discussions that we're having with yourself and others is to continue to learn kind of the outcomes that everyone's looking for, what's valuable. We're not too concerned about the UI right now. Like Emma's gonna definitely share the UI, but look at it more from a perspective of. Is it. Is it giving you the information that you need? We'll worry about the design part later. So again, I'll let Emma give you more of the details of how she typically runs these, but that's kind of the context is just a bit of a discovery and learn. On what's important on the direct and indirect bridging of forecasts. Okay. There you go. Over to you.
Me: Sounds great. So yeah I've built a small prototype and I would love to share it. But first I would like to just want to talk a little bit about the the concept. I've had a look at the file. And I've heard you're very adamant like I want this file in Palm. And I think it would be nice to just align a little bit more. I'm really curious to understand what you mean because. Correct me if wrong. To me it looks like the tab that is called. Cash flow analysis CA versus bu is the actual bridge. That you have.
Them: Yeah. Let me see. I think. It's the one I sent over to Janice. Right?
Me: I can screen share maybe easier for you to let me just do my entire screen quickly. So. This one. Do you see?
Them: Yeah. Yes. That is. What I came up with so far. Yeah.
Me: Cool. Very good. So we don't have to dive too deep into the details because I did check out your conversion table here as well. And it was. It was hard for me to follow. But let's go there later. I think it's just nice to firstly like align this is sort of what we're talking about when we're talking about bridging in your case.
Them: Okay.
Me: That correct like you have your.
Them: Yeah. I mean, it's a bridge. I don't know. Maybe it's just a table, right? It's. It's basically. Connecting both. Both. Indirect versus direct. Right. Both worlds. And obviously that's. It's sometimes. Some stuff is difficult to do. Right. So if you start up from the. The first row, three in row four, it's EVDA and EVA adjustments. Obviously, these concepts don't. Don't exist in a direct world. So these will be directly taken out of budget. Unless. The forecast inputter has a better idea because budget that's made once a year. So basically for euro ports to provide some context with your budget and then we do one or two forecasts per year. Right? That's about it. So basically it's up to the direct cash flow forecasting tool. To bridge that time gap. Right? Just to see, I guess maybe some other companies, they have a huge fpna team that runs daily sales forecast. We don't have that. Right. So it's. It's basically this, that way that we use. So ebda and EB adjustments that will have to come from either budget. Or forecast if there's a forecast available, let's say budget. And then basically anything else they can either change the ebda or eba adjustments. Or they could change row six and row. 10. Which if you change those, then it also changes the ebda, right? Because these are the two components, basically Revenue minus certain costs. You get EBDA. Right? More or less. So if you change those, I. Yeah. So that, that would be. Preferably they should, they should see, okay, budget says 5.1656. Right. If they basically say I were not going to make that, it's going to be four and a half million. They change that and. Yeah, usually they should change row 10 as well. And then basically deviation should, should, should change there. Right. This is because basically for when we're doing a board presentation, I think I also send out a sample of what the board presentation looks like. It. This is what we're so, what we're showing to the board. Is basically this is budget. But we're not showing actually, we're not showing exactly that. We're showing the delta. Right? So they know budget. The only thing want to know is how much do we deviate from budget? So it's basically column e that we show. But this is just for Belgium. Right. For Germany. So we're only going to show eventually what I'm going to show consolidated view. Right. And then what's important is working capital movement deviation and then of, of course, the other deviations here on capex side that as well. But for me specifically, what I need to answer is always. What sort of capital deviation? What's the working capital deviation? What's driving that? Is it collections speeding up? Is it collection speeding delaying slowing down? Is it. Replaying our, our suppliers too much or we have new payment term? That's, that's. That's, that's maybe 30 days and should be 60 days. What they're looking for or at least what the board is looking for is actually explanations that are cache driven. And this is the difficulty. That's why I split it out basically between. The working capital movement. If you look at 15. That split up between impact from trade receivables, impact from trade payable changes. Right. You have the movement on Gap. And then movement in working capital and for company. That they don't care about that. I cannot say it's, it's because from, from consolidated point of view. Intercompany should be zero. It should not be driving. So that's always like kind of center company.
Me: Gotcha.
Them: Line 14. That's the difficult one. What's that? That's everything that's not in trade payables or trade receivables. And that's the issue. That's the accrual basis. Right. It's other accruals. Anything accounting accrues. And I usually in budget, that's the advantage of budget. It's pretty simple budget, right? Budget doesn't usually go into detail. But I can't say, oh, yeah, working capital increased or decreased because there was a change in accounting or they did the booking because bookings don't have a cash impact. It's basically needs to be cash driven. And this is the, this is the complexity. Right. This is the, this is always the difficulty of, of bridging a world, which is booking based and not purely cash based versus a purely cash pay. So that's why, that's just how I'm doing it now. And this is, this is the initial phase. With the help of Palm. I would want to do it. Much more granular in the sense that I want to, want to split up. The section. So the first section on trade receivables and the second section, first section trade receivable, second section on trade payables. We could with Palm's capabilities of identifying transactions who those transactions belong to. We should be able to split that up saying, no, it's, it's not really Revenue in general. It's, it's, it's storage revenue or it's, it's, it's storage collections or it's thermal handling or it's, it's, you know what I mean? Much more Granular. Now it's obviously this, I cannot go into, into that detail based just on what I'm doing in excel. So what I'm, I'm hoping for is, is to get that grinder data, realize it fully that. Budget is, is a different world. Budget is less detailed, but at least from the direct point of view, I can get some answers.
Me: 100% yeah Gurjit do you want to go?
Them: Yeah. I just wanted to kind of, just from my own sake, summarize. This, this, this conversation so far. It's essentially is you have your budget, you have your direct forecast. Not everything from budget ties directly to cash. And what you're doing is you're stripping out just the cash one to one. So, you know where cash is impacting. The budget. And then from there, you want a granular view of these cash items affect these budget items and here's the drill down of why or where. You know, what's causing any sort of deviation. Is that, is that fair to say? Yeah. And then, for example, if you would look at these deltas here, right? So column E. What I would love is basically the delta is just to me to. Double click that and see, just see some transactions that are. Or something that explains it, right? Or maybe you have to, you know, the drill down capability. I have to have a full idea of how that should work, but the variances should be. Should be pointing me to somewhere, right? Or it should be fully realizing, obviously, again, the limitations of the budget that's not going to probably not be as Granular as the, as the, as the direct cash flow. But if I can find some answers in the direct cash flow, that would help. Right. It could be like, whoa, we have here storage handling, storage collection, you're 2 million. That already helps. Right. And then this customer. If I can say storage revenue, it's fine if I can say this customer, the more data I can get, which I assume is possible. Better. My answer is going to be right. Yeah. So this is, this is somewhat, I'm looking for the other ones. I mean, income tax payment management fees. Those are one to one easy capex. It's easy one to one. These are all. Pretty much okay. Finance lease. Okay. But for, for example, 20 line 27. Finance leaseless interest, I combine two budget categories because, and practice. Finance lease repayments. The capital portion plus the interest piece is going to be one transaction. So when you load up the transaction Palm, you're only going to see one transaction. There's not going to be a split. So that's why I added those up. Yeah. Interest pay. That's going to be easy. Distributions. Yeah. So most of the stuff, I think the main, the main focus is on the working capital. Which my task is basically to get the answers and usually the local countries cannot answer those answer cannot answer those questions. So this is where I'm, my main focus is on.
Me: How do you typically derive like go if we look at working capital separately? How does that process look for you today? How do you arrive at these numbers? For the movements?
Them: How, how do I get to the explanations?
Me: Yes explanations and how do you arrive at these movement numbers here in working capital today? Like what. Is your process and whatnot? Whoa this is interesting. What is your numbers come from essentially?
Them: Yeah. So the, so if you start with the revenue, that's from the, that's from the budget. That's line six. Line seven customer collections that comes from for cash analytics that comes from cache analytics. From budget side, that's a calculation. Right? Basically looking at budget revenue versus AR. If you, if you look at the formula, then you derive the customer collections. And the same thing basically also for, for payables, supply related operating costs. That's from budget. Supplier payments for cash analytics. That's from cash analytics. Or budget. That's again a formula derived between the difference between supplier operating cost and the actual movements. Movement and more capital and component Center company and the other working capital piece is basically. The whole, that space, that's anything else. And I think line 15, I don't know. I don't remember by the heart, but I think that's, that's some kind of. A summation. Right.
Me: Gotcha. Yeah. So but you're using like the direct data or basically categorized transactional data mainly to arrive at this number.
Them: Yes. So for the analytic stuff, that's all the data that you're the bank statements that they're loading into the platform.
Me: Yes.
Them: Right. That's, that's where you're essentially getting this. Like a 2.8 million or the 1.2. That's, that's actually a summation, but, like. The working capital up, it's essentially whatever your subsidiaries have loaded into. Cash analytics. And then, then you're not raising it. You're kind of just bucketing that into, like, where you think. Yeah. Yeah, exactly. Right. Except for the EBDA and EBD adjustments. Right. The, the first two lines. Anything else comes from, from. From cast analytics. And this is where then with Palm, we essentially replace that column, the cache analytics column. And that's, that would be the Palm categorized transactions. The sum of. Or. Yeah, but much more granular and automated, obviously. I mean, I think we mentioned already issues. Right? It's all manual input, manual input leads to wrong categorization. Leads to that the analysis is completely wrong sometimes. What sometimes happens is financial payments. So the, the third block, some of those payments are, are added to AP. At some operational payments, so you get a huge. Incorrect analysis. Right. Because it's not, it's not enough actual operational payment. It's a finance payment. So it's that the whole thing needs to be controlled much more because now I go through lengthy interview process, if I can call that with the local teams. And I said, oh, yeah, okay, yeah, that finance lease. We put that there. On that category in cash index and it's an it's the incorrect category. Right. But obviously, I mean, that will, that will be solved with the automatic categorization process done by Palm. Right. So that, that should already help.
Me: That's the idea I think it would be really great to keep in mind these. Indirect buckets if you want to call them that but the high level categories from your budget and how the Palm categories should map. Like what categories you do in fact need.
Them: Yeah. And those.
Me: In Granule. But I know you have a workshop coming up for the Palm categories it's just something I think. Should be.
Them: Definitely. I mean, what I, what I, the conversion table below, this is, this is just the initial view, fully realizing that cash analytics doesn't have a lot of. Categories, right? I mean, some categories, but not a lot, and some have, some of these categories are, are. Outdated. Some of the, for example, one of these categories, it's. A, there's one category called, like, cash flow from operations. Other, which is basically line 143. That's a, that's a huge bucket. That's a huge, huge bucket. Where basically.
Me: Okay. Gotcha so you need a bit of granularity there.
Them: Yeah, we need some Granarity there. I'm also working on Sam with, with accounting and it to understand how we can provide, because I think Emma, last time we spoke, you said to get this 100 right, we would need some AR AP data. Which fully honest, I won't, I won't be able to provide for all entities, not, not for the, for the, let's say, for the immature entities. I won't be able to provide that. So I'm checking to see what we can do, how we can drive that data, how we can get that data in. And I think I'll reach out to Janice or you can maybe help, help me already to see what kind of. Data ingestion capabilities that, that, that you, you support. I know it's SFTP, anything else doesn't need to be super fancy. We're, we're not, we're not a bank so we can provide the data there. For, for, for ingestion. But I have a lot of ideas on how to get this, this better. Not, not fully, not only this, but we also could add and I have to check with accounting if they have that, if they have the availability. But if they would have the availability to provide the actual balance of AR AP during the month. Then basically you could, you could also calculate DSO and DPL and then add that as a trending right during the months. You see suddenly, well, the SO is going down, is going up. What the, what the hell is going on here? Right? This, this then also helps the local teams to understand. Or maybe run rates, but that's an easy one. Right. What's the run rate collection? How much are we actually. Collecting on a daily basis, that sort of thing? Right. Is that and seeing how that trend goes up and goes down, just that sort of thing, so. But this is maybe not flee to do with this, this with this bridge here. But a lot, a lot of possibilities.
Me: For sure it's something on our radar as well.
Them: Yeah. Love the vision. I think, yeah, on the, on, on the ability to take in data. Yeah, we can be pretty flexible as well. So we'll, we'll all, I'll ping honest about this to have that discussion with you. I think as you're reading tomorrow as well. So, like, if there's time, we could talk about what some opportunities are.
Me: 100% I think probably the first like v0 of this will be naive in the sense that it will lean I'd say a on the categorized transactional data like from the bank statements to begin with. I fully agree that we need some AP and AR data in there as well. So I think that will be something that we build on top of it. Does that make sense to Matthias would that already be helpful to get like the. The bridge in a sense of what. The transactional data can provide already upfront is automated and you have the granularity and all of that.
Them: Yeah. Yeah. I think that, I think that will help because that will basically be the same thing as I'm doing now. The higher level of Granularity. In it. Right. So I think, I think that already helped. Does need to be perfect from, from day one. Right. It's. Just, just getting the, just getting the details, just getting, getting the information that maybe doing some. Some drill down, double click functionality just to see what's in there. That will be, that will definitely help.
Me: Should I actually maybe just share I were running ahead a bit like I would love to actually also have time to ask you about what decisions this helps driving internally at your ports but since we're on the topic let's just quickly jump to the prototype. And again it's a prototype very early the data will be all wrong and it's not like hey this is exactly what we're building it's something we're playing around with so just we are like first impressions. Let's see if I managed to actually share the right one. Yeah. I think this one. So okay.
Them: Okay.
Me: Imagine something like this it might be budget it might be called bridge you know don't get too hung up on this right now. But one idea is to have at the top some way. To draw your attention to where your but like direct and indirect bodies are like drifting apart.
Them: Yeah, I think that makes sense, right? It's taken into account that we're actually focusing on the working capital. That's basically be the working capital items. But I think that's correct just to have that.
Me: Could be and then you'll see this side road is popping up everywhere but you could click and then the idea is to now I've just copied some of your I've just copied a little bit but this would be Palm categories. So imagine you'd have your like. Your supplier payments bugged yet bucket and then you could. See hey what are my direct equivalents and then you could see all categories that maps to this bucket if you will. Just like how much is coming from where? Something like that is that helpful and I mean of course you can imagine further drill down from here. But is that helpful already?
Them: Yeah, I think so. And then with the drill down capability, obviously, I mean, these, these, these categories would need to be changed. Right. But.
Me: You can imagine these are like whatever you decide you want.
Them: Yeah, yeah, yeah. So, yeah, I think it would be. Yeah. And then basically having the drill down capabilities. Of, of. Maybe on the notable variance is also have already some. Information there as well. If there would be, for example, supplier payments minus 3.3 million. Would, would be worthwhile saying, oh, yeah, but we, there is a large transaction. To parsler metallops, I think. Right. This, I don't know.
Me: So you want to be able to like explain the variance.
Them: Yeah. I mean, if you see, for example, minus 3.3 million, and then basically you have structural, stable and then unfavorable variances.
Me: Or.
Them: Right on the left hand side.
Me: Yeah.
Them: I might be also useful to already have the largest transactions influencing their directly visible there.
Me: Yeah.
Them: I don't know if that's possible.
Me: Yeah.
Them: Right. And I'm not, and I'm not, I'm not looking to get it 100 to like 100,000. But what I'm usually, what I'm usually doing is. I look at, for example, coming back from the vision that I, I present to the board. Right. I start from constellation, consolidated picture. Right. So EBDA, that's all, that's all nice and fine. And then go to the working capital. And then basically it's the delta I see. Right. And then basically would be like Germany. Plus 3 million. Right. That's, that's what I need to, and then I need to. I don't need exact spin 3 million, but it could be useful to say, okay. There was something. I don't know. Unforeseen payment. Two and a half million or some or two. This, this, this is basically what I need to explain. That's the one thing. And then the other thing also I need to have those conversations with the, with the local entities presenting this view. But for their own respective entities. Where they basically say, okay, this is it. This is what it is. This is what we're seeing. And then most likely have to provide some additional color, right, on, on what is in there. The, the question is then, how do you capture those comments? I need speaking that would be here. If that's not possible, it would be me joining a down or something like that. Right.
Me: I love this feedback I have to admit an early prototype which had too much functionality in it but it did include like this like the common thing feature that you mentioned. But we can definitely dive deep into that but first things first it also sounds like the entity like it might be helpful to see this by entities somehow when you drill down. And investigate.
Them: Yes, we'll have to see it by entity. And then there's also the, the, the question what you guys are working on with the user access. And obviously, I mean, Germany cannot see this view for, they cannot see the french view. Right. This will be segregated, obviously, because access rights were actually, we, we've done quite a bit of work there, and I think we'll be, you'll be pleased with that. That's pretty much over the finish line for us now. But going back into this, going back to what Emma just mentioned, it's like what you're seeing here is a consolidated view across the whole organization. But you're, what would be helpful is if you go from consolidated to entity level and then ultimately transaction level. Right. So just three million that you're talking about. If you had a flow where you can say, oh, there's a 3.3 million difference on supplier payments. Then there's a list of the main drivers by entity. And if you click into the entity, then you see the largest transactions that are driving that variance. That essentially gets you to, hey, and we're off by three is we have this 2 million payment in germany. That was not budgeted or something like that. Right. Is that correct? Yeah. So basically it would be like, right, starting working capital. Minus 5.5. Open it up. A r AP. Other. And then basically open it, open it up further. What's the, what's driving that? Right. Or if it's, if it's directly in one of you, right? Maybe, maybe if I just show you. What I sent for Jen is.
Me: Yeah.
Them: Yeah. I mean, maybe it's more useful if I show you what I need to explain. Right. It's.
Me: Love to see.
Them: You guys have already signed an NEA. So that makes it easier. Right. Let's see. Window. This one. Share. Okay. Let me know if you can see the screen. Or not.
Me: Yes. All good.
Them: All right, so this is, for example, one of the views, right? So we usually, we usually, usually do the. Oh, yeah, I forgot to mention it's always by quarter. Right. So it's always by quarter, quarter one, quarter two. So this is basically data that we all, it's always in the same format. Right. So it's the actual, which will basically come from Palm. The last estimate, that's the latest forecast. And then usually we have like a last as a previous version. So the previous forecast. That's why I mentioned it's also important to be able to block certain forecasts because the next quarter or next time I do a forecast, maybe next month, I'll be able to be able to be, to be able to compare the previous forecast that that was communicated. Right. So that's basically, maybe this is not a good view here. This would be, for example. Right. So you see the last estimate is in blue, as in red. Then you have the previous estimate. It's in yellow. The actual, the previous year, you have the budget and then you have the actual. That's just, that's just the whole view. Consolidated view. Then we have here I've chosen on the left bottom side. Here we have actuals. We have budget just merely just purely cash position. Right. And then this is basically, that's the cash forecast versus budget. Right. So that's ebda variance, EBD adjustment variance, taxes. So taxes and management fees that could come from cash analytics. Working capital movement. That's the total deviation. And then what we usually do is split up between npl and non-MPS, MPL to give some explanation. That's the freight forwarding business. And they're all, they're always seen as separate. So we have terminals and we have freight forwarding, two different businesses. For npm in general. We don't, we don't segregate that. We don't. Drive deeper into details. Could be done in future. But then for non-MPL so terminals that we basically go down in the details saying it's really 1.3. You see all those that, all those variances. Right. And then me, I have to find out what's the reason for those variances. And I expect I, I write those down here in the text box. And then basically it's here. Right. For example, holding overdue uniforms that they needed to take France repayment over invoices, Italy. Yeah. They have an overview collection in here. I've already that sort of thing. And that's always the same thing, right? Always explaining those. Those deviations providing feedback of what's driving that increased factoring rates. That sort of thing. Right. So that, that's basically all. That's the gist of it. What I'm. What I'm, what I need to work towards, and it would be great if, if these type of views, then you're right, quarter one. Quarter two, then quarter one plus two. If I could have some kind of standby view on that. That I don't need to do any amount of work to get this into powerpoint. Right. That is basically more or less needs to be similar like this. The wording and, and maybe the view doesn't need to be exactly the same. But I get this directly out of Palm. Would help me because I don't have to manually copy in this table and this table and this table. And also the comment section, if I can already can answer that, enter that directly. Me only right now and me and Evelyn, not anybody else. I can already add this into Palm. That would be great. So this is the thing. So then this is basically. What I'm looking at. This is. So this is the board presentation. Which is one piece, and then the other piece, as I mentioned, is. There needs to be a view for the local entities. For them to view. Okay. And we just created new forecast. How does that relate to budget? They want, they need to actively see that what they're doing.
Me: Hey maybe a super silly question.
Them: Because otherwise.
Me: But is there a budget by entity? Or is the butt disclosable?
Them: It's budget by entity. Right. Well.
Me: And how does that relate to the global like you're super keen to understand how it all.
Them: So it's, it's, there's a consolidated view. Right. And then you have, you have, you have actually a budget buy entity.
Me: Yeah.
Them: But what we're focusing on is some kind of subconscious. For example, here Italy. Is actually two entities. Right holding. That's probably five, six entities. Belgium, that's six entities. And take into account that Palm is going to be like those entities that have multiple accounts, bank accounts. Right. So we'll need to get some kind of sub consolidation in place. To get this all. Similar. I can, I, I mean, we can then, that's also something to do. This is concept of groups, so we would just create groups, right? There'd be an Italy group, a holdings group, a Belgium group.
Me: 100% and then we could just make it simple to decide like if if you upload. A budget into Palm. Hey this applies to so and so group or these entities or whatever so that there's like.
Them: Yeah.
Me: Clear mapping in between not just the categories but also the entities that it applies to.
Them: And. But the only time that, obviously, this is how we present it to the board. But then again, for the entities. What I sometimes get from back from, from feedback from, from the local entities, for example, Spain has, like, multiple entities. And if I come back with the variance here, even though it's 100, 000, right? They will say for which entity is that? Right, they will, they'll, they will have to see which entity is that. It's this pain is a consolidation. And maybe they don't understand that's, but that's. And again, it's only for, for the local entities that they be able to maybe have a view where they filter on, let's say total and then maybe further down. So there also will need to be on the entity level. Right. So there'll be like a grouping accounts per entity that accounts then entities by country, more or less. Right. And then the total company.
Me: Yeah, I'm sure we could solve it like hey Spain you have access to the spain group or however right and then they can build their life on the entities and bank accounts or however they however they. Like.
Them: And then the only thing I wanted to add here, and this is maybe something temporary. What we're. Let's see. I didn't, I probably removed that. So now what we're seeing now is that, that depending on who does a forecast, their personal bias influences that. Right. And I think I mentioned before, if somebody, if somebody from AP is doing it, obviously they're over, they're overestimate payments. Because they want mail to make those payments. Right. It's, and it's, it's maybe, I wouldn't say always consciously that they're doing that, but a lot of the time it is. So what we're, what I'm, I have to do sometimes I have to add targets. I have to do an overlay. So that will basically mean. Need to find, for example, for Spain here. I don't believe they're forecast. So we'll just have to add a. We'll have to have an overlay structure adding for collections. The million at, at March. Right. And even though, I mean, once the stool will be up and running, maybe that will not be needed. But that's what I'm doing now. Right. I'm adding targets because there are some entities that are always super conservative. And prudent in their, in their estimations. Or maybe they don't know.
Me: How do you currently have these conversations like is there is there any tool currently that you have access to?
Them: Right.
Me: That can help them like I guess I guess it's not that easy so you know.
Them: No, what I'm sending is basically what I sent to you, right? The Excel is what I'm sending now. And it was already, it's already cleanup version from what I said before. And I asked them to provide feedback. Right. And some entities will be. Like, like Finland will be, they're more complacent. So they'll see deviation versus budget. I'll just adapt the forecast and maybe I'll match budget. Spain will be. It's the business. That's always the answer. It's the business. It's because. But you don't understand this is the business. You know, it's always depending on, on the, who's ever in the seat, but the tool needs to be able to provide them. And basically also expect. That's why the kpis will be super useful. Saying, this is what we're seeing. That's why I'm trying to get the three years that that jan has asked for. Right. This is what we're seeing in the history. This is what we're, this is what we're seeing here. Right, to also see this is what you're doing usually. Right? I think it's never going to be perfect. Right. And, and, and, and it's always the human aspect you cannot remove.
Me: For sure.
Them: But there needs, needs to educate not only us as treasury. But also, also then show them what they're, what they do and how, how, how their business is behaving. Right. It's a, it's much more than it's purely a cash flow forecasting. It's also an educational tool, at least for the local entities.
Me: This makes a lot of sense. So essentially the I don't want to point fingers but let's say the Spanish entity they could access this tool in Palm and they could see for themselves.
Them: They can see for themselves, Seymour versus budget. And they should be, they should realize if there's a big gap.
Me: Yes.
Them: As narrator. Right. And then also once we have the, the kpis in place. You'd say, okay. You're clear. You're clearly underestimating factor seats by 50.
Me: Look at the history.
Them: Talking about this and then we'll push these forward. And once it's in Palm, we'll just blast it out on a monthly basis and I'll put everybody in copy not to shame everybody, but it's, it's the way you should. You know, you should say, okay, these, these entities will present to the CFO these entities. These are doing a good job. These entities are not doing a good job. To provide some at least some, some gentle guidance. There they should see the kpis themselves as well. But, you know, sometimes it's, it's, it's, it's how you say that's the carrot and the stick. So, you know. What you're saying is you want a big scoreboard that when the CFO logs in, it says which entity is then the most accuracy? Yeah. And then this was actually one of the things you mentioned, right? He wants to know which entity they're going are doing because basically I'll send it out. And if there's no response, and obviously you have to go over their heads and it's not, this is not what I want to aim to do. But sometimes that needs to be done. So.
Me: Yeah, fair enough. No but I hope we can ultimately deliver something that and also functions it's a collaborative surface like at the end of the day as well as much as like it drives decisions not only at group level right but it drives decisions at the operational level in terms of how you work together. And that that will roll up. So I think. That is a good insight.
Them: Yeah, yeah, yeah, yeah. So I now have had discussions already today with a guy from it just to see how we can get the data meetings on Friday with. With accounting to see how we, how, how everything flows, how we can get certain, certain islands, because as I mentioned, if I would want to split up. The trade receivable piece in a trade payables piece. With more data, then I would have to be able to quantify that. Based on the, the, the profit and loss statement where you have some kind of. Segregation of the revenues, right? You have segregated storage, this is, this is. I don't know, this is. This is straightforwarding, this type of thing, right? You get that split. So let's see how, how, how we can get that data. How can, how we can Define certain rules or we can. We can look at the transaction data coming into Palm. And say this customer is, for example, storage revenue. Right? Or it's not going to be perfect because obviously you will have customers that, that, that. Buy more than one service. Right. And then in that case, you'll have to. Make a best guess assumption. And it's always useful to say, okay. We're doing good in search revenue, right? Coming from this customer, I'm just basically saying. Collections are going up. Right.
Me: Of course. And I guess sometimes your biggest customers can drive pretty big impact as well it's enough that one of your customers. Changed their behavior or something and then you'll see the impact I'm just guessing. But.
Them: Yeah. That, that could be. Well, yeah, I mean, obviously we have a factoring program for most identities, so everything, everything is factored, right? Basically means that even if they pay too late, we get the funds. To a certain limit. If they pay over 90 days, then this gets cut out. Of the factoring program, meaning we, the financing that we get up front. Gets penalized because of the overdose. So it's a different process versus regular customer collections, which, which we do have. But. Let's say 50 from a global perspective, 50 is factored. Right. Everything in Europe is factored. Everything outside of Europe. It's, it's not factored huge, usually because, yeah, factors want to have the. The easiest jurisdictions to do the factoring. Right.
Me: Cher. How do you currently track like. What how much of the factoring comes from customer if that question makes sense?
Them: So honestly, we don't.
Me: Okay.
Them: Right? We don't, we don't have the issue is. And this is also at some point when I use Palm for as well. You should the data with that, the factoring, the factors providing, it's cumbersome. It's a lot of, so a lot of stuff is just in PDF, so I have to do like run a script on it or I use, I don't know, power query to get that out. It's, it's, it's. The only thing work. For example, this is, this is also what we present, factoring percentage. But that's basically by entity. So we don't have any more data. Would be great to see. Okay. If you can double click here and see, okay, why is, I don't know, why is this entrop here? 83%. Oh, yeah. And then double click or something and say this is mainly driving by because our sugar revenue with this customer is, is increasing something like that. Right. To get. That, to get, because now obviously, I don't know if these tables are not, not that user friendly, right? It's more, it's probably better just to get some kind of. Three, four or five bullet points saying, this is what it is. And if you want to see anything more information, look at the appendix. But, you know, get, get more of the data out of it. But then. We'll have to see how, how, how we handle that, because obviously. From a transactional point of view. When, when the entity. Gets the payout from the factor. The data doesn't mention which entity, right? It just says factoring payment. So we need to have, and as I already asked to the factor as well, we need to have access to their platform. Or some way to get. To get to see which customers pay. Right. That's the otherwise they'll just be factoring percentage increase without any additional value. Hey, Matthias. So I'm going to try to play back, like, the factoring Journey just to, again, get it right in my head. To do business with the customer. You have invoices, you sell them to your factory bank. I think this is ABN or one of them. Yeah. Avian and Amsterdam. Is it just others or is it just abn? It's just abm. Cool. So then there's some sort of process in which all these invoices are bucketed, sold to abn. That's what's going to build out this percentage number is how much of your invoices are factored. Is that correct? What you're showing here. And then every month or whatever the Cadence is, ABN pays you a sum. Which is a consolidated whatever agreed amount is based on the invoices that you sold to them. Not every moment. Right. It's basically on, on every time you, every time we upload invoices in the morning. Yep. They review it. Deduct any payments that are received and they do the payout. So it's, it should be on a daily basis that we get funds, not just on the monthly basis. So it's a, it's a daily, it's a daily basis that we get. Got it. Okay. And then, so the part, the problem there right now is you get a lump sum from abn, but that doesn't really tell you, like, which of the customers. Have. Well, it doesn't matter if the customers are paid, right? Because you're selling the invoice and you're getting the discounted cost of that. So where the, the concept you mentioned around the sugar company. They're, they receive, the receivables for you guys are increasing because of X, Y and Z. Therefore your percentage of factoring is going to go up. How would you, how would you determine that? Because you would have to then see the factoring. Invoke. Is it the invoices? Is it the sell, the, the sale of the, the invoices to the bank? Like, at what point do you recognize? Oh, okay. The reason it's going up is because. Well, what you would need to see is, on the one hand, you will need to have the AR data. Right? And on the other hand, you will need to have the. The fact and payouts to determine. I mean, look at this formula as here on the top left. Right. So basically, factoring advances divided by the trade receivables. So you need, we'll need to have the same thing. But then on the granular level. Right. Yep. So the trade receivers, obviously, that we can get from, from the ERP. Right. With sftp. Yep. You guys load that up or we can load it up and you ingest it. But then on the factoring advances, I have to see maybe I can also. Add data might also be available through because accounting needs to reconcile it. Right. The question is just how, what's the best way to do that? And that's why I have this meeting with. With accounting this week and it just to see how, how that's done. Do we need, do we need to build a link between Palm and ABN? Or is the data already there? And is it already get closed when you, when you, I would imagine it does, right? Once you send the, the factoring over to ABN. So, like, hey, I have a receivable from company a for a million bucks. We're going to factor away 800k today. Receivable is for whatever 30 days from now, 60 days from now. When you, when you go ahead and factor that now, ABN has it, they pay you back for that in your ERP. Is that, that receivable didn't closed so you don't, you don't have it as an AR for 30 days out. It's essentially settled. No, how it works is basically. Or at least not works at our company. So outstanding, outstanding AR or trade receivables. Will be there as long as the customer didn't make the pay. Ment. Right. So you have basically, let's say you have 100 million of trade receivables. You'll have. 75 million of factoring advances. And you have a lump sum of SPV. Basically, we, we have agreeing with the factor that the factor advances 95% And 5% is on the SPV side. Which gets released once the customer makes the payment. So imagine that the customer pays on 30 days. We upload the, the invoices they won. We get paid out maybe day one, day two. Right. So that factor in advances shows what's already collected. The, the traders equals will be in there as long as the custom doesn't pay. Once the customer pays, then basically the factoring advances gets counter booked. The trade receiver gets counter booked and the SPV as well. So that's, that's how. Okay. I think for me, I'm getting, I get the confusion there is like, is there a double counting in that case then? Because you're getting your, your, your funds from the factor, then you still have open, the open receivables. Yeah, but the, the factoring advances a negative sign. Right. So it's, it's going to be on the accounting sector. Okay, so you're off booking the cash something. You're booking it at something else. Basically. It's part of the trade receivables booking, right? Yeah. So you have total total, you have, you'll have total trade receivables, and then you'll see trade receivables. Right minus factoring advances. And then you have also the spv. So we're negative sign. Got it. Got it. Okay, that makes sense. And it should net to zero. The, the three of. Yeah. If you have 100 million and 75 millions in advanced, then you have 25 million s net trade receivable. That's how, that's how they, that's how they do. It. Okay. Cool. So I'm hoping I can get some, some stuff out of, out of their system. If not, we'll need to get, we'll need to get access to. To ABN, which I know ABN as an API, but that's for the bank. So I'll have to know if ABN has also an API connection then for. For refractoring division or not. So that hopefully, hopefully it will be based on an internal data.
Me: Maybe it's too nitty-gritty but just like if you have. This daily process where you decide what invoices you send over for factoring. How does that process look like? And who does that?
Them: So. It's AR or credit control, depending on the entity. They, they upload, they upload data from the ERP with tools called the basics. It's a tool that it's required requested from abn. Also, I think it's also a Dutch company, the basics to, to do the upload.
Me: Okay.
Them: And then I think there's some filter in there. Right. So probably we cannot upload it to company, right? If a company cannot fund, we cannot upload, needs to be clearly netted from payments. So there's some kind of filter on that, and it's, it's sent over to, it's send over to alien. Then ABN in the morning, they will check, right? See if there's anything. Anything that needs to be corrected. Anything, for example, overdue if you get over plus 90 days, you get penalized. Right. So they will reduce that. What else could be? And that's the, that's the anointing. That's the annoying piece with, with fact and versus securitization. Factoring is, is always credit limit dependent. So we have a line straight, which is our credit insurer. If they provide a limit of 100,000 and we're trying to get funding for 150, then we're also only get a hundred thousand. So 50 will not be funded. Right. So it's not always going to be like everything we upload is going to be accepted. No. But still, I mean, from the booking process, it'll be. Anything that's acceptable with the fact in advance. Anything that's not acceptable just, you know, be considered as straight receivable. S.
Me: Gotcha. Yeah now it would be I think the easy the most straightforward just in my gut feeling would be if there's a way to get the actual data directly from them what was being paid out like what invoices.
Them: Yeah, that's why I think as well. But. I first want to see what, with, with it, what they can do themselves.
Me: Yes.
Them: But then still, I mean.
Me: I was just curious yeah, I was just curious about the like uploading process because that in itself contains the invoices that should map to a batch payment from a young so I was just but yeah like you say it's not always that.
Them: It's definitely going to be correct. I mean, it's never going to be. Yeah, the only, the, the only. Sort of, source of Truth is going to be actual payouts. So that's going to be. And then the other thing, just to make it more complicated in a sense is that we're using the terminals are using Oracle. And then freight forwarding is new is using, is using the vision. So probably it's going to be easier if you get. The data directly from abnormal.
Me: Yeah.
Them: You have one source. That me having to, having to set up one connection for one, let's say one sfp connection or whatever for, for your Oracle and another one for, for in the vision. Right. I mean, there are also some other small ones as well. So I think the AVN kind of confirmation, if they have one that says, hey, we've accepted these, these invoices is, is would be like the, the magic bullet, and at least in terms of forecasting and kind of capturing the, the cash received, I think you still face a problem of your trade receivables that are not accepted and what are still outstanding. Because. ABN might tell you which ones are not going to accept, but we don't know until what's, what's in the ERP, right? Is at some point, I'm assuming it's fed back in. But I don't know, I think it's something. Yeah, we could, we could spend time on that. And when we get there and try to see if there's some creative ways to get that information is really, really interesting. This is going to be a fun one. Yeah. I mean, for me, it would make sense, right? If you have the transaction information and then you have on the one hand the opening R. The only question is then. Will you be able to 100 match those payouts. Versus customer fear versus invoice level or not? That's. Maybe, maybe that does not, does need to be invoice level. Right. But at least it should be customer level. Because then we can ask that question like that we're, that we're asking here. That sugar, sugar customer here. 99 factor. Great. And this, this also comes back into all the question of what's driving that working capital. Right. Is it just a very good, this is an easy one. For, for easy explanation for working capital. I mean.
Me: Exactly.
Them: Tackling percentage is decreased. Why this customer here? Why invoice is not approved or anyone this, that sort of thing, right? It's, there's not, this is, this is an easy one. This is going to be easier to explain than explain why. Why we're, why there's a deviation on the payable side. It's, it's. Right. Man. I really hope we get this data. Because I think I, I do think I'm pretty optimistic on if we do get these data sources that the capability to do some of this matching the commentary is, is highly like, like, do I do think I'm optimistic it's doable. I think the, the problem is not the problem, the challenge is, is the data in itself. And, like, can we get it into the platform to then be able to do the analysis and commentary? Yeah. And I should get access to that, to that ABN platform soon. It's usually for accountants, but I requested, like, access three months ago, but they send the, the access token to send the Luxembourg and have to resend it. So once I can access, I'll probably see what it looks like. So. Okay. Because I was thinking, obviously it would be, it would be great to have that direct connection, maybe API or whatever they offer. If they don't offer it. I don't know what you can, what, what, what, what if you can do assume it's, assuming it's an easy login, right? Which from what I'm using, the other platform of abn, it's just username and password. If that would be the case, maybe you can scrape it somewhere. Right. If I provide you with Username and access, you could scrape some of the data. Yeah. Yeah. You'll, you'll hear that from Giannis tomorrow, probably is that we, we are building that capability or testing that capability for not just these types of portals, but also Banks. So if you're comfortable giving read only access to, like, a dummy user, we should be able to then automate how we pull, pull data. But TBD on that one, just keep going back in your mind. I'm sure you guys will bring it up tomorrow. Okay, cool. I think, I think, at least from this perspective, I feel pretty good. Emma, did you want to show anything else more on, on the Prototype? I know we're, like, right at.
Me: Yes. I think this is more than enough validation for us to keep going in this direction I just want to again highlight I think v0 is going to be like what we can read from the transaction categorize but I think it will be something that already will help you answer like we want to help answer as many questions as possible obviously about what's driving the variances between your direct and indirect budgets that's kind of how we're framing it like the v0 the variances.
Them: But.
Me: What's driving them? And then yeah we'll work from there into the depth of this feature. I think.
Them: Yeah, no, no, that, that, that, that sounds great. That sounds. We'll help. We'll save a lot of work for me. Reprogram the Excel errors I'm having. So there you go. There you go. I think, and we build from there, right? So for tomorrow, this categorization meeting that you're having with Giannis, I'll also ping him. But I think it looks like we'll have to do. Bank categorization. So statement categorization or cash categorization. And then also the bridge over to what are these categories? What are they tight? Like, which category do they tie into in your indirect forecast? Yeah.
Me: It's a good thing back of mine I'd say for tomorrow.
Them: Yeah. Yeah. Yeah. I mean, yeah, maybe you don't do the mapping tomorrow, but, like, keep that in the back of your mind. Yeah. The full mapping. I still need to, I still need to have that meeting with it this week as well. But, yeah, when I met, when I met in Antwerp, we did discuss the, the concept. I think we already had it for Belgium. Right? Now that it's 100 correct in red category versus direct category, we already had that based. So I think he's, he's aware of the, of the, of where, where I'm going to go with that. One final question on this and assume at the end of the, at the end of the process, the ultimate goal will be then to have, you know, the chatbot, asking the question and obey the answer that on the bridge as well. Right. Not that it's a must, but it's nice to have. Yeah, absolutely.
Me: I was thinking that when you were talking about your your charts that you needed and your commentary like that sounds like something we could definitely leverage AI to help create based on your actual data of course. So the feature has a few like different dimensions it has an ingestion part like how do we get the data in? Palm? And then there needs to be some sort of persistent version of it like this this is your your bridge and then of course yes 100% super super relevant to be able to chat to that data and perform like more analysis on top of it.
Them: Yeah. Yeah. I'm even thinking through, like, even now, potentially, say you get one of your entities with a bridge, with the category, the bank, the direct and indirect bridge, the category alignment. I think we can already test whether the chat bot can already address some of that now without an actual interface. Right. If we load, load the mapping into, into pulse, basically.
Me: For sure.
Them: But.
Me: It would yeah it would it would be temporary for the session.
Them: Yeah. Yeah, yeah, exactly. Exactly. Just, just, just to see if it works on. And then ultimately, I mean, basically, if this, this, if this question here, right, this, this, this text box could be partly at least answered automatically. Not that those will also be useful. Right now, that's a must. But again, it could basically say, I don't know, these are major influences, and it's probably not going to be always going to be correct, but at least there's some data in there because it knows it. Right? Yeah, I, I guess I'm getting ahead of myself. Like an idea. I was thinking through is if you have a section and Palm that's giving you all of the commentary that you can maybe select pieces of commentary that you want to then transfer into a report. Right. So here's the 20 things that have happened. But for you specifically, Matthias, you're like, well, these are the five things I actually want to communicate. The other ones are maybe more so for me. Click these five, push those into my report. And when it publishes, make sure that this commentary is included. Right. So I think there's a lot of creative ways where we can leverage AI, the large language models to do the commentary and then we figure out a nice way to insert it into your reports when you're publishing them, I think could be pretty cool. Yeah. Yeah. And then something similar like that, right? Something like for, for sure, the graph on the quarter basis, this table, it needs to be similar. I cannot, I cannot create like a new completely new revamped board deck needs to be similar, right? It needs to be, as we're seeing here now. So sure. Yeah. Yeah. All right. Very good. Very interesting session.
Me: Thank you it was really a really good talk I'm sure we will actually we'll talk more about this but this was really good directional feedback for us so thank you.
Them: Right. Yeah. I think, again, as Emma mentioned, like, we're, we're, we're gonna spend more time on this one, and hopefully we can make some more time over the coming weeks. But the goal is try to get this into the product ASAP. Right, Emma? It's going to be in there ASAB. And so if you're willing, yeah, I would love to catch up again in maybe in a week or a week and a half just to share, like, kind of the progress and, and then eventually push this in for testing. Yeah, I'll send you, I'll send you, both of you. I'll send some, you some feedback based on the discussion I'm having. On Friday. With, with it and accounting. To see what, to see what they say. That you can already see where we're thinking about it. And then, yeah, we can do follow-up follow-up session and then see what's possible. Awesome.
Me: Sounds great.
Them: Awesome. All right. Have a good one, man. Yeah. Thanks, guys. Bye.