Name: Smarter Cash, Sharper Decisions: How AI Transforms Treasury Management
Uploaded: 2026-04-23T14:46:13.576Z
Duration: 1 h 5 min 42 s
Description: Smarter Cash, Sharper Decisions: How AI Transforms Treasury Management

Transcript for "Smarter Cash, Sharper Decisions: How AI Transforms Treasury Management": Good morning from Detroit. Anthony, great to have you here. Everybody else, let us know where you're dialing in from. I know we've got a massive audience. I think we broke the four figure sign up number. Yasmin, good to have you from London. I'm from Greece right now. I'm in Athens. So, yeah, let us know where you're you're dialing in from. Jens, good to have you. Singapore, Miami, Toronto, keep them coming. Let's try and get South America in there. Good. Good stuff. A lot of Atlanta. The Americans coming in full force. Quebec, n y n y. I like it. Guatemala, Central America. Good stuff. Great to have everybody here. I can see we've already got a 150 people in the room. Let's just keep waiting, for a couple more minutes. While we do, I'll introduce myself. My name's Nat. I'm your host today from AduCap. I've been with AduCap, for about three years, both as head of partnerships and, an account executive. So a demo of the tool, monitoring its evolution, as we continue to add in AI features. I'm joined, today by my colleague, Kelly, who is our AI engineer at building many of those features. I'll let her introduce herself shortly once Nicola joins us as well. We'll just wait a couple more minutes, while, people, continue to join. As we do that, I'm also gonna open up, our first poll of the day. So I don't know what it's like in in in your jobs, but everyone, seems to be, coming up with a new, AI use case every single day, at AdigiCap. So how much do you guys use a AI? Do you wanna just quickly describe your own, personal experience with AI? I should have opened up the poll now, so you should be able to see that. In the polls tab, if you look to the right hand side, you should be able to see just next to the messages area, you should be able to see, the polls. So if you guys fill in that poll, we'd be really interested to hear how much you guys are using AI, what tools you know, are you using multiple tools? Are you just interested? Do you feel like you should be using more AI but haven't, had a chance, to get involved because you're just too tied down with all the tasks that you have to do on a day to day? Help us fill in that poll, please. Alright. So, Nicola, great to have you join us. Hello, Nicola. We were worried. you weren't gonna rock up, Yeah. but it's. good good to have. you. I I thought that the I had just a small point too, so sorry about that. You probably got, yeah, bogged down in, some AI workflow that you're gonna tell us about, I hope. Anyway, welcome to the stage. You've arrived just in time. We have, just released our first poll, but I think this is a good opportunity for you to introduce yourself to our audience. So, Nikolai, obviously, one of the I think you've got over 1,000,000 followers on LinkedIn now. As you, might have been doing just now, you've obviously built and tested, you know, lots of AI workflows. But, crucially, I think for this audience, you've been a CFO yourself, for many years. You've been through many of the pain points that they've been through, and, you know, you know, requests from the CEO last minute on a Friday. So can you tell us, Nicolas, having just joined now, having been a CFO for so many years and been through, so many of of the pain points, what made you walk away from being a CFO, which is obviously a great career path that many people on this call would really wanna become CFOs to really focus on AI. It's a bit of a you did it four years ago. I think we ourselves did our first webinar three years ago. What made you make that move? And when you look at our audience here today, what would be your advice on how to use AI as a real lever without feeling like they have to change everything at once? And, yeah, what advice. would you give? So, so first, thank you, everybody, and and sorry again for the delay. I revalue the time of everybody, so, we'll catch you back on this. But, really important is now when AI arrived, I saw a huge opportunity that I don't think a lot of finance people saw three years ago when you arrived. And when I was working in finance, one of my favorite part was helping the others, like giving them Excel tips, ERP tips, SAP, PowerPoint. And I found that it was where I was the best at helping other grow rather than just doing pure finance task. And so for me, it became really a mission. I started on LinkedIn. I grew, really fast having, like, also a lot of cheat sheets that I guess a lot of people saw. But when AI arrived, I saw really straight away the potential. And when I was talking to around me, like, everybody was talking about how AI is going to help marketing and help a lot of, tasks, but not finance. And, so really quickly, I will I just, like, made my mission to show that to people. And today, that's what we are going to show. It's really going super fast, but, actually, it's much more approachable than what you think. And that's what I want today to share with everybody because the tools that are available today, you can all have them and do so many great things, that I'm sure you are not, aware of because there's a lot of things that you don't know that you don't know. And so, I know that for me, this is a mission, and, I want to also show practical stuff because I think there are a lot of webinars where people only talk. So that's why I'm excited to start and show use cases as soon as possible. Yeah. I think that's so true. I I find, at least in my own experience, getting you know, using AI has been a bit like writing an essay. You know, the top the hardest thing is just getting started and testing. And then, afterwards, it's about where to focus, which hopefully, well, I know you'll be able to help everybody here with, today. So I think it's time for our second poll. For those of us, who missed the first one, don't worry. We've got you covered. So I'm just gonna close the first poll and open our second poll, which is to what extent is AI currently used in your workplace for finance tasks. I have opened that second poll. Tell us, how much AI your team is using right now, by filling in the poll. We had, I see, over a 150 votes for the first one. So, really, let's try and get to 200 in this one. Really keen to hear how much AI there is, in your company. If there's any skepticism for the companies you worked at, also really curious to see that. Let us know, guys. Keep, filling that in. And then let me just say a little bit about AGGAP. Obviously, Nicola, we want to get to the workflows that you'll hear to show our audience today, but, obviously, it would be remiss of me, to not say something about AGGAP, the company hosting, this webinar. For those of you that don't know us, ATCAP is a next generation treasury management solution for the mid market companies. So we are really trying to fill this gap, for treasury solutions for finance teams, that are operating, you know, below kind of 300,000,000 in in turnover size, where you might be wearing many hats. You might not have a dedicated treasury team. We're trying to provide a solution that automates many of the treasury tasks, including via AI, for finance teams like that. We were born in France back in 2016. We have since expanded, to have offices in, I think, seven different countries, US, Spain, Portugal. I'm usually in London, Berlin, Italy. So seven different countries, and we've got over 8,000 customers in 2026. So that's a little bit about us. We're ranked by g two, which is a software comparison site, as the leader for both treasury management systems in, the mid market, and, also for cash flow forecasting, which is, our our bread and butter was our original platform. Excuse me if that's a bit showy off y, but I think our marketing team would kill me if I didn't mention those two badges at the bottom. In terms of what our tool does, I like to think about AGGAP as kind of a one stop shop for cash. So we have kind of two core objectives. We want to centralize all of our customers' cash inflows and outflows in one place to give them visibility of their cash today and, in the future, both medium term, short term, and long term. And then second objective is then to use that data that we've centralized to help our customers increase their cash efficiency. And to do that, you know, not only does that mean a more accurate forecast, and we'll talk later on, about how we do that, but, also these dedicated modules that we have to get cash in faster, the accounts receivable automation, where you can identify your late paying customers, come up with automated workflows to encourage them to pay more promptly, and really sync all that activity that you do to to collect money, with your cash flow forecast directly. And you can do the same on the cash outside where EdgeCap offers customers both the ability to, manage, supply invoice management, supply invoice workflows, so verifications of the invoices, etcetera, and then also build payment files directly in Anchicap and send them to the bank completing all the approvals in Anchicap, and basically execute your payments without ever having to leave the tool. So no more bank logins to complete have a final approval. You complete all the approval stages in Anchicap, then just send the file to the bank who then executes the payments on your behalf. So that's our one stop shop for cash. We integrate with full range of, ERPs, NetSuite, QuickBooks, would be the most common ones in The US, Business Central, all of the the main ERP providers you can think of, we do integrate with. To kind of put that into a clearer picture of different roles within the finance team, I usually find this slide can be quite helpful. Basically, if you think about splitting the function into these kind of four subdivisions, AGGAP, ultimately, the treasury systems, we're very heavily, focused on this. We support heavily, and and Kelly will show this in some of the AI areas we're really focused on, in accounting and operational tasks. So credit control, like I just mentioned there, supply payments, and then also a a key one for finance teams, I think, the bank reconciliation task where AI can be really, really strong, and then kind of classic reporting tasks, as it relates to cash flow, and doing budgets. Aducap also supports finance teams in that regard. So that's a quick intro to us. I don't wanna take up, any more time. If you do wanna learn more about Aducap, you should see, at the top of your screen a request in Aducap demo. If you think some of what we offer, could be relevant or you're just curious to learn more about how we, how our cash flow forecasting is different from, other providers or, how we can help on payments. You know, can we handle particular payment use cases, which banks we work with, which ERPs we work with? Just request an AnchCap demo, and you could set up, a call with myself or one of my colleagues to learn more about AnchCap going forwards. But without further ado, I think, Nicolette, it's time for me to hand over to you. So I'm gonna stop sharing, and I'll let you share your screen. Yes. So, we'll start with, one of the use cases is, reconciliation because I know so many people, work on this and lose a lot of time. And let me show you what you can do already today. So this here is yes. So the first use case, you imagine that I know for finance, a lot of people spend a lot of time, like, organizing their invoices. And before you can even reconcile, you want them to have clean in your, in your, organization folders. This is already what you can do today. So on the left side, I have CloudCoWork. On the right side, I have the invoices that are not organized. Like, it's. just, like, 100 of invoices. I want to have all of them organized by currency and by month. But for this, I will need, like, one hour to open each of them and then save them at the right folder. So instead of doing that, I will give cowork access to my folder, and I will ask, so called cowork, to, organize these invoices into a month and into, currency. So also really important, when you give access, you also have the risk that it can delete the invoices, so always have a backup. So here, we are just asking in plain English, and this is where also those AI tools get better and better because they have more access, and you don't need to have, like, a complicated prompt for this. You just need to know what it can do. So just with the sonnet model, it can do it. And now you can see that I'm going to launch this, and all of the 100 invoices will be reorganized. So before doing my reconciliation, I just want to have all of this, cleaned up. And we can see on the right side, for now, they are not organized. But really quickly, we'll start moving. And instead of me spending, like, one hour organizing all of my invoices, I will have CloudCoWork that open each of the invoice, read each of the invoice, and now starts to create, like, a folder for January, one for February, one for March. And for each of them also, USD and Euro folder. Now you can see it on the right side, and now it's done. And if after I open one of the folders to verify so let's open maybe, the folder, January 1. So we can see on the left what it did. So it organized all of the invoices by month and by currency. And in January, if I open one of them just to spot check, I can see that this invoice is from January and in euro. And I can open another one to verify. And this is where it can do that, but it can also, like, create Excel file. It can also create a lot of things. So this is a a simple use case just to show you that it can work on your computer. But the real work is in the reconciliation. So I want to ask in the chat, like, how many hours today are you spending on reconciliations? So if in the chat you can swear, that would be great. Too many from Katya, two days. Yeah. So let's imagine now, this bank statement. And I have this bank statement in Excel, but I want to reconcile it with all of the invoices that we have organized before. So in the same discussion, I just say, now reconcile these invoices to the bank statement, and I will upload the bank statement. And we will see that Cloud cowork will call something which is called reconciliation skill. What's really important is that Cloud has these skills which are embedded and that you can call. And thanks to this skill, it knows how a reconciliation should look like. So this is in the background doing the work. And after a few minutes, which I accelerated a bit, it's generating the working file. And if you think about the reconciliation normally on paper or, like, on, in Excel, it never looks good. It's not easy to understand. Here, I want everybody to observe how this reconciliation looks like. It's super neat, super clear. It has classified our 100 invoices against the bank statement. I know what I need to check as a reviewer. I also know the reconciliating the reconciling items. I know why. And I can just go over the full reconciliation and finish the work as a human. This, if you want to do that manually, this is hours, or even like I saw a lot of people in the chat, days of work. And this is possible right now with these tools, but I will say this is possible when you still have something manageable, like, in an Excel file with, like, hundreds of lines. I I know people that have tried it with many more lines or many more invoices, and these tools have limits, with this. So it's quite good if you are a small companies and you have not a lot of lines to reconcile. Once you need to do this at scale every day or to connect with your ERP, to connect with your bank, I will say this is also the limit of these tools, but this is what AI today can do with generic tools. And, maybe Kelly will can explain later, but the limit is it's not connected yet, and also you don't have that many lines that you can process. So that's like you can see I did it with 100 lines. So that's the the the difference between, like, a specific tool and a generic tool like Claude and a specific tool like AGCap. Yeah. It's phenomenally powerful from a from a data mining perspective, and I agree with you. The UX is just, incredible if you think how long you would spend just doing the formatting, to kind of look like that after having done reconciliation work through. Amazing. But, Nicola, that's obviously kind of a a very data heavy task. You know, reconciliation is about it's a matching exercise. Right? What about where we have to combine that kind of data analysis like Claude just did with visuals, like in a board pack? Yeah. So one thing, I wanted to present as well is we lose so much time on creating a PowerPoint presentation. And I'm curious in the chat, like, let us know, like, how long you need to create a PowerPoint presentation for your, month's end or for your weekly review. Because I know from my past, like, I needed, two days, like, also with a lot of work around it, but just in PowerPoint, at least half a day. Yeah. I see people confirm that, like, two hours, three hours. Well, now let's imagine if your working file is done in Excel. So, again, let me just do this. So you have here a really clean Excel file where you have done your working. And here, because my file is well arranged, normally to go to PowerPoint, I will need to do a lot of, storytelling, a lot of copy and paste. But here, I will open cloud in PowerPoint. I will just mention this Excel file. So I'm explaining that I'm a CFO. I want to create a board presentation based on this file. And because the file is well structured, I don't need to give that much context. And that's a big difference with before where, before you had to prompt really, really well. You had to to give a lot of details. Now these tools understand how to make board presentations. And so with just uploading the file and saying, I want a board presentation and also giving my branding guideline, I can have PowerPoint a code in PowerPoint working on it, so thinking about it. And now looking at the file, just asking me some questions. So what is the audience? What is the size of the deck? So what do you want to have? What is the visual tone? And once you have answered to these questions, it can go ahead and proceed on the work that needs to be done. So now you can think again. Like, I think a lot of people say three hours, two days, four hours. Within like, now it took, like, five minutes, which accelerated a bit. And I confirmed that I'm okay with the structure because on the right side, it gives me the structure, and now it says, let's go. Let's let's build this PowerPoint. And now you can see, like, the first page there. This is an easy one because it's just the presentation page. But the second one has, my KPIs, so my executive summary, as you can see now. And then it will build all of the other ones with the right storytelling, the right graphs, the right details. And just about, like it took me because now I accelerated a bit for for the use case because I don't want everybody to have to just steer steer as a at a at cloud, but it took around, like, seven to eight minutes to build. And you have a full board pack finished also with the risk and ops, with the cash flow. And now your presentation is done. Of course, after, as a good manager, you need to review. But again, a tip for everybody, don't review manually. You can ask AI to do the first pass of review for you, to verify again that against the Excel file, there is nothing which, is wrong or hallucinated or misunderstanding. And then you review it last time, and then your presentation is done. Wow. It really is, phenomenal. Yeah. I think of, so many times as a consultant, spending so much time, doing all this analysis, and then you have to spend another day putting into PowerPoint. So eight minutes is a sort of a a a scary timeline, but also a a great relief, I'm sure, to the next generation of consultants. I guess, one of the themes that's emerging here is, you know, when we look at that presentation, you know, we have to trust the data's perfect. You say, you know, you're gonna ask AI to review it, then you're gonna review it. Many people, I'm sure, on this call will be going, wow. Our data's in a state. How are we gonna, we we're not gonna be able to take advantage of these tools. Can AI help them out with that as well? Yeah. I saw actually, John, saying that in the chat line. Is there a way to better structure? Well, the the good news is here, cloud is not available only in PowerPoint, but also in Excel. And maybe for the people that might be frustrated thinking, oh, yeah. But I don't have cloud. I have Copilot or I have Chargegpt or I have Gemini. Well, good news. You have also Copilot in Excel and Chargegpt in Excel, and both, they can do what Cloud can do. But, really, in you're going to see now, in 2026, if you are not using AI in Excel, you are still in 2025. It's really important that it'll be like, today when somebody is drafting a long text, you always start with AI because it's faster and better. Here is the same, and I'm going to show you, why. So here in terms of data cleaning, I have these two bank statements where for each bank statement, I have, like, three years of history. But for each month, it's in one tab. As you can see, when we open, for example, this one, you have one tab per month. And so until I can do forecasting on this, until I can do a nice presentation, I have first to consolidate and to clean because also, the banks have different formats. The date is not the same. So I'm going to open Claude in Excel. I'm going to add these two files, and there I can ask Claude to help me consolidate these two files into one, but not with 36 tabs, with also just one tab with all of the data consolidated. Like this, I can do a pivot table. I can do an analysis, on it. I can do graphs. I can do, formulas with only one, table. And so as you can see, now my data is consolidated on the second tab. And also really important because I see so many people saying, how do you know it's correct? Well, I'm always asking now in my prompt, make sure to audit that your work is correct. That basically, the single tabs and the consolidated tabs is the same. And that's why we have this audit tab that I've asked to be created to see if there is no mistake. And now I'm also starting to clean the data so that it looks better, that I can reuse it. And as you can see, now I start cleaning it to have a better tab, which I can work with after for a presentation. And you can see also on the date, the format of the date isn't the same. So I'm going to ask for the date to be clean, etcetera etcetera. So I will spend a year some time cleaning the data, but not anymore, I will say, manually, but with AI. And I think the last part was, to clean a bit the date. And you can see here the audit. So we have audited that all of the transactions individually, they are the same than in the consolidated tab. A big recommendation because we are still in finance, meaning that you cannot trust AI if you didn't review it. You cannot trust AI if you don't have a process to make AI audit itself and to have a documentation of what happened. But once you have this, it's basically like somebody working for you. When you take over the work, you want to make sure that there was a review done before, that there was check, that there are controls. Here's the same. We are asking AI to do these checks, after. Awesome. And once you have done that, what you can do, just to continue, you can ask AI to help you build a forecasting model if you have, the cash flow history. So I can call again Claude, and I can ask here, k. Can you build, like, a simple, cash flow forecast model? So it's not going to be a crazy AI with, like, machine learning, but just something simple to start with with formulas, with assumptions that can change. And to come back to my point at the beginning, if today you are starting from scratch without having AI helping you to start building the model, you are still in 2025. Today, AI can do in a few minutes a model which is like, it looks nicer, right, than what we can do. And, the formulas are all there. You have also here assumptions you can change, and, it will probably do a better job than you. But, again, we have to review. We have to edit that there is no hallucination, that the, logic is correct. And then once you have done that, then you can use it. But the review is much faster than, doing the draft by yourself. So I'm curious what people think in the chat. Yeah. Keep the questions coming in the q and a, guys, because we wanna we will make time for the q and a at the end. I know we've got quite a lot more to get through, before we get to that. But, yeah, it definitely leaves us wondering kind of, and curious to get your take on this when we get there, Nicola, kind of where the role, for finance professionals is in the future, you know, because it feels like, the people who have the domain expertise and know how to build the models have a huge advantage. You know? And what about young finance professionals? Where should they focus their energy as well? Would be I'd be curious to understand that. So, Kelly, I think it's time for me to hand over to you. I I guess it's me to let you share screen as well, and you can leave the slides. Yep. On our side at ADICAP, this the first feature I want to show you is categorization with AI because this is one of the most important AI feature we have in ADICAP. Categorizing every incoming transaction is a huge work, really time consuming. So this feature really helps our clients, with this task. So the idea here is that you have all your incoming transactions, out inflows and outflows, and this AI algorithm will help you, find the most important word in the group of transactions. And by most important, I mean, the the word that holds the most semantic sense in the title of the transaction. And from this word, the AI will suggest a rule that you can see here on the left. So here the rule is when, the transaction is cash inflow and the title contains the word client, no matter the bank account or the or if it's a paid transaction or a future transaction, I suggest that you set the category to one of services. And if I decide to apply this rule, what's, wonderful is that it will not only categorize every transactions, of this group that I can see here, but also every future coming transaction that will match this rule. So it's a huge time saver for categorizing every transactions. And then so here you can just adapt the rules and everything is created. Then I will let you, Nate, talk about our reconciliation module that is also really awesome for this. Nice. So thank you, Kelly. What Kelly was showing you guys there is, the AGI Cap expected tab. So in AGCap, when we do cash flow forecasting, we have not two data points like you would have in Excel, which is your forecast, and then you actualize this forecast. We have this third data point, which Kelly was showing you there, the expected transactions. And this is because we want off our cash flow forecast to be based on the real operations happening in the business. So the expected is my, customer invoices, my purchase orders, all of these, items for which we have real data feeding directly into AchiCap, from our ERP system, and from any other systems that we might have, like payroll or, our CRM system, feeding into AchiCap so we have up to date evolving view of cash flow based on direct data feeds, from other parts in the business. So what it means is ATCAP wants to be a source of truth for two things, both your upcoming view and your current view. And your current view is your bank data, and we connect in AGICAP directly to your banks. So we pull your bank statements into, AGICAP directly, and we have your expected transactions, which is, all of the future bank payments we might consider them. AKA, it's a reconciliation task. We want to reconcile our bank data on the left against our expected transaction data on the right. This has been a task that finance teams have been doing for a long time, obviously, since, double entry bookkeeping was invented. And even with spreadsheets, it's always been very, very time consuming, and it's done by most of the finance teams that I talk to, no matter how much they might be using Claude or or ChatGPT in their personal lives, still manually. And what AchiCab wants to do is for those higher data volumes, so if you're beyond us SMB, is really start to automate the matching. And it starts to get quite interesting when you have scenarios like, you know, a customer who's got 10 outstanding invoices or maybe 15 outstanding invoices, and then they pay 10 of them in one go. You can start to have that single batch transaction mapped against many, transactions automatically via AI because the AI can recognize the customer name. So, again, like Kelly explained, AGCap will look hard at the data here that's in the name, will look hard at, you know, which bank account it's due to go into, which category it's put into, what date it's it's, due to come in on or did come on come in on in the case of the left hand data versus what when it was due to come in, on the right hand side here. And we can, in AGCap, then build various rules to really begin to automate the reconciliation as much as possible. So you can start to have invoice and credit number recognition, you can have customer name recognition, You can adjust for days in the future. You can adjust by value. So, you know, if a customer ends up paying you £999 instead of a thousand pounds, Adjacap will auto reconcile. So you can really start to build all these different types of automatic reconciliation to then build in that main view, which I showed here. It first a suggestions tab, which you as a human then manually net off, and then here, a manual tab. And do you have some kind of security around what you do? Aducab retains a history around exactly the behavior, and then the goal is to create a bank journal file that we can then match against all of your general ledger categories and feed in to your ERP to create, to to help you with your bank reconciliation. That's the goal. So we serve as a source of truth for your banking data, source of truth for your upcoming, forecast data or your expected transactions. We match these off to keep your forecast really accurate, and then we use that exercise that we're doing to avoid duplication in the ERP, and then we send automatically the data back into your ERP automatically mapped to your general ledger codes so that your ERP remains a source of truth. We do this on an ongoing basis. So that's kind of what we're doing from a, AI perspective when it comes to reconciliation. We're using ourselves as a source of truth and building these kind of rules, on a customer level, a category level, a date level, an amount level to then ensure that the AI can make suggestions, and then ultimately automation. And we're starting to see, bringing in an agent. It's gonna execute this as well, which is currently running its own tests on one to one transactions, and the next step will be having the agent run on one to many transactions as well. So that's kind of what we're doing from a reconciliation perspective. I think, Kelly, I pass back over to you now to continue, with the other initiatives you're running at Edgecam. Yes. So Nicolas showed you many use cases in cloud, and it appears that those AI agents are becoming more and more common. So, at ADCAP, we are working really hard on developing our MCP. So what's an MCP? It's basically a connector that will allow your AI agent to read your ADICAP datas, to analyze your ADICAP datas, but also to take actions for you in ADCAP. So it will help you automating your work on ADCAP. One of the stakes of the MCP is is also interconnectivity. The idea is to connect your tools with other tools. So here is an example of how you can do that with Agika. The use case is a use case that has been asked by our treasurer, that is sending every first of each month a message to our CEO to inform him about the cash burn of the previous month. So here, I'm asking Claude to prepare a message. I I give it, the template with all the numbers I usually send to my CEOs. So what Cloud will do is it will fetch all the necessary datas from my ADCAP account, and it will run many calculation as to collect every number I asked for. And it will start drafting a message directly in my Slack for my CEO. So here on the video, it has found all the numbers of the figures, and it has drafted the message. I can see it directly in Slack. I would be able to review it, eventually edit it, and send it. So this is awesome, but we can go even further because I will ask Claude to schedule this task for every first of each month. So then I won't have to ask it again every first of each month. It will fetch the new figures of the past month and prepare and send the message in my behalf. So this is something you can do with Cloud CO Works. So here, the task is being scheduled, and then I can review it to ensure it fits my needs. So here we are. The task is scheduled. I can see it in the schedule list. I can see all the instructions, so all the prompt, and check the the schedule, so every one of the month at 9AM. This is a really interesting example of how, the MCP will be will help you connect your ADCAP datas to your daily workflow even when they need other tools like Slack. So, this is the project we are working really hard on to to release really soon, but we are also working on other project including, how we can improve the forecast in ADICAP. So the first thing we've done to help you, improve your forecast is detect recurring transactions in your history. So we have an algorithm that will look at every past transactions you've uploaded in ADCAP and detect patterns to make suggestions of transaction you think should be created, in a recurring manner in the future. So it will pre fill everything for you to review and create as a recurring expected transaction. So in the future, for example, every month, on the five of the month, every day, etcetera. This is really helpful for your rent, your fees that are coming regularly so that you don't don't have to think about creating those. This will really help to improve your forecast accuracy as those transaction are already, already in Agika. And then once we have done that, we want you to go even further, to help you with your forecast. So we wanted to try, and forecast your cash flow based on your history with AI, like a super machine learning model that will, help you forecast your cash flow. It is not yet available within IGCAP directly, so we decided that we will start implementing it through the MCP. So for this example, I asked Claude, to call this, this super IGCAP model to plot accumulative daily inflow line chart for one of my category that is operational flows. It's a it's a super parent category with several subcategories. We will see the date the detail later. But the idea is just, okay, call the super ADCAP machine learning model and show me what can be the future, basically. So here, could have many data to retrieve, like the categories, the the past data to show me, on the graph, and the future data from the supermodel. And then it will run this as a cumulative chart as I've asked. So you will see in blue what's, what's the past, and in green, you see the prediction with the light green being a confidence interval so that you can trust the the AI output. And then as I as I told you, what's really powerful is that we can, decide what granularity we want for this forecast. So I can decide what category, what subcategory I want to forecast. So here, I ask Claude to deep dive into the subcategories of this, operational flow categories. And you can see this is the same graph as we saw before. And here, I can see categories by categories, the forecasted cash flow. So really useful for long term forecast and having an overview of the future. And this is for this is it for the the use cases I wanted to show you in AddiCap. Wow. It's so powerful. I mean, Claude is just phenomenal for for visuals. So yeah. and this is what a lot of people are asking the how can I get access to the data also in my AI tools? I was, like, two hours ago with a on a one to one with a CFO. They were asking that, like, how can I get my data inside also the tools I use everyday? And the MCP is really the the solution for this. Exactly. So the vision we have at EdgeScab is that we have EdgeScab as your source of truth for cash, and you use the MCP, to do your analysis in the way that, Nicola and Kelly have been showing you. Okay. I think it's time for our our final poll. So I just, reshare my screen, and then I wanna get, to the q and a. So one more poll, guys. If we have, if I understand correctly, the poll is after seeing the use cases, how do you currently feel, about oh, I don't know why it's in I have to close this poll, and then I can open this one. After seeing the use cases, how do you currently feel about the increasing role of AI in finance? Are you guys excited? Are you guys a bit uneasy? I've seen some nervousness about security on the poll, on on the chat. So are you nervous about your own jobs? What's the what's the feeling having watched this, today? Curious to hear from all of you. Let us know. Last bit, and then I wanna get into the q and a. ATCAP's own position on where we stand on AI. Finance teams are not unique in, being affected by AI. This applies to sales teams. This applies to marketing teams. I think, you know, seeing those visuals, you see how powerful, tool this is. But we're big believers that AI is not gonna replace people, that it's an enabler. From a a finance team perspective, we believe, as Nicolas showed, AI can improve data quality and, therefore, help finance teams, get better quality data in faster time. We believe it can help all finance teams, if they use a tool like AchiCam, of course, to centralize all of their cash data and then AI tools to analyze it, improve their cash performance, and obviously use that information to predict with confidence and ultimately reduce costs. You know, if you have better cash flow, you need less reliance on external financing. You can better optimize the money that you do have, both in terms of investing in, and pushing it around your business. So in terms of what we apply at AGGAP ourselves, we run a program where we deploy AI where it creates the most value. We have a means of, you know, a company wide initiative of, AI testing, and then we pick and choose the best initiatives. We'll try to make all of these transparent and all usable by design so that, you know, whatever AI testing gets done can be repeated so you don't waste hours, going down a chat, tunnel, and then finding that you have to go down another chat tunnel to repeat the same task. It's very important to avoid that by building projects. And auditability is obviously critical. We know that AI is hallucinate. You've gotta remain in control of what you're doing, and, usually, that means having, clear traceability and clear responsibility, on individuals. In terms of what it means for treasury, which is obviously our domain, we think, mass market AI tools such as ChatGPT, Copilot, Claud, I know someone asked on the chat, you know, can Gemini, ChatGPT, and Copilot, do the same thing? I think Nikola was pretty clear that most of them can. These are gonna become, more and more widespread. Their use cases are only gonna go up, and so, you need to have a structured adoption, to get the most out of these tools. And the key to doing so is kind of having a centralized data source, which you refer to for particular tasks. So, obviously, from a sales perspective, that'd be your CRM. From a treasury perspective, we hope it'd be a tool like AchiCap. Best practices, we will recommend based on, working with our clients thus far and what we're applying within our own business. You know, structure your prompt prompts very, very clearly. I think this is something, I tried and tested myself with Claude a lot. If you're very clear, it it massively speeds up the process of getting the outputs that you want. Definitely go into subtopics. I think, you know, when you come at it with a massive prompt, somehow there's more room to get lost. Active training, I think I saw some stat the other day, you know, that, you know, 9090% of teams want to upscale, but only 4% of people are actively upskilling. As a company, you need to drive that or as a as a leader within your team. If you start, people will follow. Data integrity is at the heart of it. You know? All of these AI tools are only as powerful as the data that they have access to. That means strict governance standards are critical, and then don't get rid of the the human validation, method. Even if, as Nicolas showed in his example, he asks, AI to produce, the document audit, someone needs to audit that. I'm gonna stop there. As I can see, we're getting to the ten minute mark, and I wanna leave, enough times enough time for questions. But reminder, if you wanna have a look about what Agicap does, and learn more about us, just request an Agicap demo at the top of your screen right there. But if not, let's get over to the q and a. I see some people have been posting a lot. Nicola, I don't know if you've been glancing at it, while I was speaking. Are there any questions that really stood out to you that you think we should answer straight away? I mean, there is the big question about data confidentiality. So I think it's really good that we think about it again. For a lot of years between the arrival of cloud tools and now, like, being just used to you to use a SaaS, you know, like software as a service online. We don't even think about it anymore. Like, I know so many people use Google Translate to put a really confidential document into Google Translate without thinking about it. Now people are thinking again that they need need to pay attention because AI seems more intelligent. And this is good to think about it, but it's not good to stop because of it. So, yes, you need to pay attention. No. You cannot stop. Like, do not stop because, actually, all of these tools, if you get a business license, then they are under SOC two type two, compliance report, which is the best standards in terms of, data security. And especially these SOC two type two reports were created by in The US by the AICPA, so the, the organization, which is basically like the the account and organization. So it's a standard which is recognized. And normally, unless you are in the defense industry, in the health industry, in the banking, this is, like, the standard that should be okay for your business. You still need to need to know if you have higher, standards, but this is like a common standard. And if you take a business license, then you know that, Cloud, JGPT, Copilot, they are at this level. So what happen often, however, is that employees or consultants, they will use an individual license, like, Chargebeegee Pro or Plus or whatever, what is called, Pro Plus, Free, Max. This is not business license. And those one, they are not protected by a sub two type two, report. And there, you are putting yourself in not in a good position. So that's why if you are a CFO or if you're a manager, make sure everybody is doing is using AI with the credit license and also give it to your team. Because if not, they are going to use AI with their own license and that you don't want as well. Yeah. I think that's a critical point. Right? People are gonna use it anyway, so, make sure, that you've given them access and then then it's controlled. Interesting question I saw from Paul Cross there. Do you think it's a winner takes all market? He's saying, you know, he remembers Lotus one two three, but Excel, won the battle. Do you think all of these LLMs will continue to survive, or do you think, will have, one or two that dominate? So for me, I I say that since, I think since two years, I see Microsoft winning the corporate battle, because, ultimately, they have distribution everywhere. They have, like, everybody is using Microsoft for Office, for Outlook, for Azure, for Power BI. So they're everywhere. And they are also, you can use the models of OpenAI and Opus with just, like, the the Microsoft front end, which, by the way, like, everything I showed to you, now it's already available in Microsoft if you use, for example, Microsoft CoWork. So it it's like you need to have a frontier access, but you can activate for you can activate it. You can have Copilot in Excel, which does this as good as what we just did with Cloud in Excel. It's just people don't know about it. So Microsoft, for me, will win the corporate war. I think then, the individual, you know, or, like, small companies, then it's still not sure. I thought for a while that JPG will build a new Apple. I see that now cloud, is catching up, but we never know. So let's see as well because, there is, like, a lot of money needed to run all of these LLMs. And maybe if one is not there anymore, it's not because they were not performing, but because they didn't have money anymore. So let's see. Okay. Edward's got an interesting, point here, which is that, it's gone a bit fast, for him. I mean, if he wants examples, I guess, he better head to your website, Nicola. That's right. Yeah. Or to the so on YouTube, I have much longer, videos about it. They are free. And if you are serious, you can also come to the AI Finance Club. It's the number one community, on AI for finance. We are 2,700 AI CFOs, and there you get, like, unlimited access and live master class. Is the right place if you want to learn and you are serious about it? Okay. Nice. Any other questions that we should definitely answer? I'm just trying to there's been so many. It's quite difficult to scan through them all. Yeah. Maybe this is an interesting question. You know, how long does it take, someone who hasn't used, AI but is obviously very familiar with Excel and PowerPoint to do the research and learn these tools and get up to speed? I mean, the problem with AI is there's a lot of things you don't know that you don't know. A big advantage right now in this webinar, it goes fast, but at least now you know it's possible. So you will know that you have to learn it. But there are a lot of things you don't know that you don't know, and you need to put the work on it. That's why for me, it takes, like, ten hours per day since three years just doing that. But that's also why we created this community so you have in one place everything you need to know about AI for finance. I also think that if you are already somebody who is curious and who asked the right question and who knows that it's important, you will allocate the time the the time. A bit like if you wanted to run a marathon, you will put the hours every week to run, and to do the miles, or the kilometers. Here is the same if you allocate your time, I'll say, like, two to three hours per week to learn. Plus if everything now you do, you try to do it with AI, you will learn super fast. But there's a lot of, topics, for free. It's a bit unorganized everywhere online, or you can go to a a dedicated platform like ours. And then I guess the the obvious question is, it's clear that, you know, a lot of, finance work, should we say, is gonna be automated. You know, tasks that used to take eight hours now take one, or tasks that used to take, days, might take, you know, a couple of hours. Where should finance people, you know, invest their time in upskilling that's not AI tools? You know? Where's the con where's the value gonna be on the on the human touch, going forward? Yeah. So for me, I always relate to two different type of type of finance people. You have finance people, they are really comfortable behind an Excel file all day long. And, you know, also, like, how since COVID, like, a lot of people stay home and don't want to go into the office. The thing is, this is really at risk because you can see how I could do the this Excel model super fast. But if you are the one, talking, to your colleagues at the coffee machine, if you run through the the floor of the production and you come across a machine which is still packed, and you know because you are following your CapEx project that this machine is actually normally started to be activating this month. And you know all of what it represent in terms of cash flow, in terms of depreciation. And, normally, your mind should go and ask the person next to it, say, why is it still packed, and why are you not using it? And then the person will say, oh, we are still missing missing missing one piece or certification before we can, use it. They would tell you it takes another six months. And then as a good finance person, you should go and change your forecast straight away. And, also, if it's really like, if if it has a significant impact, also escalate. this info super fast. But this, you will never have this info if you stay behind your laptop and your Excel file. And this is going to be a big change between letting AI do the thing that are have no human value and you going and do even more human task, doing also more than what the management and the business needs from finance. Meaning, each time you see a problem, you don't just observe it and wait. You go and you find a solution for it. Each time, you go in a board meeting, you already have, like, questions pre answered or ready actions to solve the gap between the budget and the actuals. Each time, AI is also evolving, you are making AI doing this mundane task and not just waiting that somebody else is going to program that for you. So this is the finance team that will win. The finance team finance team that will not win is the one that will just ignore all of this and keep doing the same work and work and work, and then slowly, they will lose a bit of, of time. Okay. So it's a bit of power to the extroverts and a bit of power to the problem solvers. I guess, the people that go above and beyond, Yeah. and recognize where, you know yeah. It tends to be true that, you know, once you solve one problem, another problem emerges. So, I think it's just gonna be a case of this you're gonna be able to get more done in a day and the person who makes themself valuable will make. themself valuable. Okay. We're at time. Nicola, are you okay to overrun? Yeah. Okay. So let's pick another question. Are there any other, points that you saw come through in the chat that you wanted to, definitely answer? There was a question from Zero. Right? Maybe it's more for you or for Katie. Zero energy cap. Does it integrate with Xero? Yeah. There was a question for this. Like, if I'm using zero, how does it work? Do I import the reconciliated statements? For Xero, we have two way, API integration. So, for your forecast, we would connect directly with Xero to pull in your, invoices, purchase orders, etcetera. And then, from a reconciliation perspective, you then wanna feed your bank journal data back into Xero. To complete the reconciliation, we would send this data back, on a daily basis or or whatever frequency you need. Let me just consult the q and a again. I think this is an important one. Can you help our audience understand, you know, the differences between Claude's chat, Claude's co work, and Claude's, code? I guess Claude Code is maybe a bit more in the name. But how and when should they use them differently? Yeah. So I can share my screen. It's the easiest, I think. So let me share. Can you stop sharing your screen, not so I can go and can't. share? Let me stop sharing. Yeah. There we go. So let's put it here. So here I am on the web, and the web, this is what we called cloud chat. Okay. It's just like the a bit like chatty bitty. Like, you just talk and say, like, And then you just talk with this, and then, it will reply to you, which has a lot of functionalities. You can search web. It has projects and all of this. But so this is normal chat. Then you can download cloud on your laptop. And so here, if I go to cloud, so now I am and, let me show we can start with or let's go here. So now here, I am with Claude. So I can have the chat, but I can also have co work, and I can have code. So the chat is the same chat as what we had before, but what is the big difference is co work because co work can, work on your laptop, and so have access to your folders, create things. But also CloudCore can take over your Chrome browser. And just like for for the anecdote, I don't know if I still have it, but I was just before with my team, we did, like, a workshop, and we had a a picture of, I don't know if I still have it. We had a picture, you know, like, of what we decided to do. And I just asked Claude to make this in a FigJam, so like a digital board, and it took over my Chrome to make it. And, I don't know if I have it here, but I didn't touch anything, and it was able to do this. Like, just like doing all of this by replicating what was on the on the picture. And that's like the beauty of it with CoWork. It can access to your computer and really do things. You could basically say, oh, go on QuickBooks and fill up, based on the invoice, fill up the journal entry. It can do that because it can take over your computer. But it comes also with risk because you need to know that when it takes over the computer, you need to know what you give access to and what you don't want to give access to. Awesome. Thanks, Nicola. I think that's a great way to, end, our webinar with an extra demo, so to speak, and a key distinction between, Claude chat and Claude Koa. Thanks so much, for coming on, the show today. Reminder, for those of you who wanna learn more about Adjacap, just request a demo, or get in touch with me on LinkedIn as well. Very happy to answer any questions you may have there. Nicola, how should people best get, in touch with you? Go to your website. Yeah. So, again, like, I will paste the YouTube channel to continue to learn. And if people are really serious and want to learn, we have the AI finance club. We are 2,700 members growing and having the best and smartest people, in the world on AI for finance. So if it's for you, if you feel like you want to learn and become one, you can join us. Great. Awesome. Thanks so much, Nicolette. Thank you, Kelly, for showing your initiatives that you're running, and we'll see you all next time, guys. Thank you so much. And here, you. Thank you. the webinar. Bye bye. Bye. Bye.