Transcript for "European Customer Panel: AI Won't Fix Finance - But It Will Change How You Work": For CPE credit, be sure to stay through the full session and respond to all poll questions. Welcome back to day two of Take Control. To kick things off, a fresh perspective on how AI is impacting global finance teams. This is AI won't fix finance, but it will change how you work. Kicking off this session, Hugh O'Neil, principal accountant in residence at FloQast. Hello, everybody, and welcome to day two of FloQast Taking Trail twenty twenty six. Really excited to be here talking to the panelists, the esteemed colleagues I have with me today. And my name is Hugh O'Neill. I'm the Principal Accountant in Residence of FloQast. And today we're going to talk about the topic of AI, surprise, surprise. But what we're going to focus on today is the reality of what's happening. So we've got three European leaders of finance teams here. And we're going to talk about how AI won't fix your finance processes, but it will change how you work. I think we've seen a lot of change in the industry over the last few years with the advent of AI, the change it's brought on. About 77% of finance teams now are using AI in some shape or form, but we really want to talk to three leading practitioners here who've seen some practical use cases and ways that they can improve the operations of their finance team. So without further ado, I'm going to introduce Nicola, Diane and Steve. I'm going to ask them to give a quick background of themselves, a little bit of an overview of the company and the team they're working at and their journey and how they use FloQast. But, Nicola, let's start with you. Yeah. So I'm Nicola Rowlands. I'm the deputy CFO at Dept. Dept is a digital marketing and technology services services partner working across the full customer journey spectrum with our clients, with a team of 4,000 Devsters based across the world from Australia through India, across EMEA, and into The US. As for finance, I work... Well, our finance department is split between four pillars. We have client finance, FP and A, operational excellence, and then financial control, reporting, governance, and compliance, which is what I mainly oversee, including the finance products, which we work on, which is NetSuite and, of course, FloQast. Brilliant. Thanks, Nicola. And it's... Yeah. It's a really large international group and lots of different entities, lots of different legal structures, lots of detail for us to get into during the conversation. Brilliant. Diane? Hi. Yeah. Diane, I do, finance systems and change chapter lead at Tesco Mobile and leading the finance transformation program. I've been at Tesco Mobile for six years, and Tesco Mobile is one of the largest MVNO in The UK. We are a joint venture between Virgin Media o two and Tesco, and we have, a fairly lean finance team. We've always been very manual in our processes, and what we're focusing on is sort of kind of growing that team and kind of improving on the efficiency of it. In terms of what we use FloQast for today, at the moment, we're using the compliance module, and we're trying to standardize our documentation, understand where our current processes are so that we can track how we then start to progress as we transform and manage any kind of gaps and deficiencies. We're also looking at the month end close to all, which we're hoping to implement soon as well. Brilliant. So change, I think, Everybody's talking about AI, but I don't think enough people are talking about the change that it's bringing. So it's gonna be really interesting to get your perspective as we go through the conversation. And last but not least, Steve. Hi, Hugh. Great to see you again. So I'm Steve Bourne. I'm a senior director at Michael Page. So I run the global records report team, the purchase to pay team, and what we call the global finance systems hub, which helps to manage our finance applications as well as increasingly change automation and AI in finance. So I've got two hats. I run some of the operational teams, and I'm driving the changes in those teams as well. So Michael McPage is one of the world's leading recruitment firms. We operate globally. We specialize in permanent interim contract recruitment solutions as well as executive search and global outsourcing solutions as well. We contracted pay... We contracted FloQast five years ago now, I think, originally for the balance sheet reconciliation solution, which has worked brilliantly for us. And since then, we started using balance sheet variance analysis. We use it for our month end close checklist. We also use it for our local compliance tracking, which is where we make sure people are doing things like VAT submissions on time. We've just bought the AI matching capability, which looks very exciting. Brilliant. Thanks very much. And I think it's great that we have the range of perspectives here as well. Diana, you mentioned compliance as well. And Steve, you mentioned compliance. Like when we were talking and Mike went through his keynote yesterday, he went back what he called the ABCs of finance, but making sure everything is done accurately, which I think is really important. Everything is compliant and then the B in the middle of it was making sure everything is you know, within budget. So how we get that accuracy, how we get that compliance, but keep managing the operations and the effectiveness of the finance team, I think that's going to be really interesting as we go through the conversation. So we're going to take a pause here and and take the first poll question. Let's jump into our first poll question. Where would you honestly place your organization on the AI adoption curve? Brilliant. So now that we're back, we want to talk a little bit about, the reason why teams are adopting AI. And there's in my many years as an accountant, I've never come across somebody who doesn't want to close faster, who doesn't want information quicker. But I think in the middle of all this technology and all this pressure to adopt and all this change that we're going through, people are kind of not they're losing why they're putting this technology in the first place. I suppose what was the driving reason... Nicola, we'll start with you. What was the driving reason between you looking to adopt this technology and and bring it into place at Depp? Well, I suppose Depp is just pioneering at heart. And when we see, you know, the business doing so much stuff with I... AI for our clients and just, implementing stuff on their on their side of things, they're driving us also to do what we can on the finance side of things. And, yeah, just automation and doing things more and better to better quality and standards is just always our drive. And AI is now enabling us to do that Yeah. Across the team. Steve, I know this is something... I wanted to come to you because I saw you give a presentation at Gartner a couple of years ago when you talked about the journey you've been on at Michael Page. Talk to me about some of the projects and kind of the the journey you've been on over the years implementing AI solutions. Yeah. Sure. I mean, there... There's certainly pressure at the moment. I'll talk a little bit about the recruitment industry, first of all. So there's two main challenges that we're facing. One's cyclical, so when the economy is buoyant. Times are good in recruitment, and we saw that after COVID, where we had two fantastic years as lots of people changed jobs. Since then, we've been navigating more turbulent economic conditions, and that's put a lot of pressure on headcount and cost saving. We've seen a bit like Nicola, the front office, the business starting to take in... Initiatives in the AI space, and we've been asked to do the same. I think the business trusts us to take sensible decisions, and also create a a platform that the business itself can leverage in the future. So we're doing that in different ways. We've been acquiring platforms such as FloQast to know exactly what they're doing and make our life easy for us. We've also been exploring our own agentic AI functionality. And just recently, we deployed our first solution in our O2C space where we use AgenTic AI to perform compliance checks on our revenue processes. That's really interesting. And I think that balance, and we're going to come on to that later in the conversation, buy versus build. But kind of seeing the range of solutions and the range of options that are right there, I think it's really important that everybody really understands the reason why the different solutions that they have. I suppose, Diane, what about you? What did you... What was kind of prompting this this change in the evolution of how you're reporting at Tesco Mobile? Yeah. I think there's probably two drivers. While we're still early on in our AI journey, I think what we're focusing on is getting the foundations in place so that we can enable it in the most, efficient way and not not have to go back and retrofit anything. The biggest driver's boss has been sort of risk. We have a lot... Because we're the finance team, we tend to have a lot of single points of failure. And so we are thinking about how do we make our processes more resilient. And then the other is we're an agile organization. And as part of that, finance needs to be able to adapt to the business and be able to support agile decision making. So it is how do we get information that's more reliable, quicker, faster, and thinking about how do we then kinda keep pace for the business as they transform and ensure that we are fit for the future as well. And one of the things I wanted to touch on here as well is is, you know, what's driving this change? Are you seeing this change come from leadership? I hear anecdotally from a lot of people that every CEO is selling every CFO to think about AI. Were you seeing this kind of change driven by the team as well? So as Nicola, where are you seeing this change coming from? I mean, it's both. Leadership obviously do want us to do it and make the changes that they want us to make, but it's also from the team. The team are also pioneering at heart. It's in the DNA of depth to make sure that we're automating and and constantly improving. So, yeah, it's both sides, and the team really want to embrace it. And we actually did a survey recently, and that turned out that the team were actually really excited about AI. I thought it would come back. There'd be a few nerves about it, but, actually, it was overwhelmingly excitement about how it's gonna change their roles. Diane, from you, you know, you introduced yourself with with that change, but, you know, that changed title. Like, what... What's really driving this change for you at Tesco Mobile? Is it, like Nicola said, kind of a combination of everybody wanting to to try and do things differently? I think the main drivers are probably the finance leadership team. So we recognize where the rest are. We recognize where the opportunities lie. And so we've been kind of steering that by defining what our long term strategy is and where we want to get to within the next five years. The dealership team are really supportive, and I think another driver for us is making sure Tesco Mobile is a great place to work. So are we using are we using the best of our skills and abilities, to do our everyday job? Do people have job satisfaction? And so a big part of this is making sure that we're an organization that attracts good talent, entertains it. And you can't... Like you said, everyone's using AI and they're personalized. So to not use it in your in your work life is is almost an injustice. Yeah. There's definitely both pull and push. There's a really strong pull, from our CEO and CFO. They constantly talk about the importance of automation and AI and being technology enabled, and we've seen them invest. Right? We've seen them invest in tools such as FloQast, recently the agentic AI solution. It wasn't cheap, but they invested. We bought other applications, and they've fortunately invested in my GEFS Hub team, as well this year, which is great news for all of us. And then there's a push from our teams. As soon as the ladies here, we... I think we hire certain profiles that don't wanna do routine manual work. Their transformation polls... Profiles, they see opportunities. They attend panels and conferences like we do. So there's a lot of drive coming from our teams themselves. So one of the things as well, and Diane, I'm gonna start here with you because you've mentioned risk so many times. But one of the things that we wanna talk about as well and focus on, especially in the European context is working across lots of jurisdictions and lots of different areas and lots of different teams. And the setup that we see, you know, will bring complexity, and there's a different regulatory view here in Europe on AI than there is in The US and in China. I suppose how are you dealing with this risk in in that profile and when you're thinking about, you know, the jurisdictions and the and the places you operate in? Yes. I think from a complexity perspective, Tesco Mobile is a UK business. So we don't have the complexities of cross border, like regulation. But I think for us, it's more sort of making sure that we align to both our shareholders, security considerations, cyber risk, and all of that. So there's a lot of, red tape, you would say, for us to navigate, really understand, and make sure that we are completely compliant. So making sure from everything from, you know, procurement process all the way through to delivery, we have done everything we needed to, followed all of the guardrails to make sure that whatever we are implementing is secure. Yeah. And I think I think that stakeholder management and communication is really key, especially, working in a jurisdiction like that. And, Steve, I remember, when I first met, Michael Page, you talked about the international operations and the shared services. I think somebody said it to me that there's wonderful Buenos Aires, Barcelona, I think it was somewhere in the East Singapore and Slough in The U. K. But that's a real like multicultural mix of different people. How do you manage your team in the kind of that complex global environment? Yes. So we're seven fifty people in finance now. We operate across 40 countries and mainly through our three shared service centers. They're now in Buenos Aires, Barcelona, and in Kuala Lumpur. We're from Singapore and and Slough. How do we manage standardisation and consistency across the teams? Well, we have global process owners like myself. So every single week, I talk to my global leaders. We have global town halls. We have global risk and control matrices as well. We try and have global systems wherever possible. It's not always that easy. We have global standard operating procedures as well, and we have a global transformation team which works with SMEs in each of the shared service centers. And that sounds like, Dan, you could probably talk all day about this, about a lot of stakeholder management there. But in that global context, things aren't always the same and they will be different and you'd expect them to be different. How do you manage that risk profile when you're dealing with, you know, acceptable changes versus following the global policy? The answer is it's not easy. I think we all know that automation and AI work best where data processes and rules are the same. So where we can, we standardize those. So for example, we we implemented, an expense audit solution using AI recently. The only way we could work that working... Get that working effectively was by creating a global expense policy, which we didn't have with global thresholds and, expense allowances on that as well. That works, but in the Adjentic OTC, AI solution, we had a greater challenge. So I'll talk about that. So in Michael Page, we recognize revenue when both the candidate and the client confirm a placement. That used to mean we had lots of specialists, trawling through lots of documentation, emails, and other documents to ensure that things like the start date, the salary, the job title were all agreed. Now we have agents doing that, so the agents will sort through all of the material, and then it will make a rep... Recommendation to our specialists. So that sounds great, on paper, but in practice, we found lots of procedure exceptions. And actually what we found was AI wasn't the complexity. AI surfaced complexity. So some of the differences we saw were what actually, is effective as acknowledgment. So in some countries, it might be a job offer being signed. In others, it might be a contract. In some countries, we could use documents. In some countries, we could use emails. So we try to standardize where possible, but ultimately, we had to create our own country specific agents for those tailored processes. I think that's really interesting coming back to what we're focused on here is it's going to change the way you work with servicing that complexity is really interesting because people think and coming back to what we said previously, like the CEO will just get change from seven days to three days. It isn't that hard, Steve. Just, you know, they don't... And they don't understand the kind of the complexity and the nuance and the detail. Did you see something as well, Nicola, I suppose, when you you operate in a global... In a global... In a very global business as well. But as that consistency comes in like Steve is talking about, does that actually say, oh, actually we need to take a step back and think through some of these processes? So we have a very structured onboarding process. So we've been a buy and build strategy as well at depth. So in the last ten years since it formed, we've bought 33 entities, 10 in the last ten in the last five years as well. And that just brings complexity. And, again, the the one thing that we say is nonnegotiable is everybody has to be onboarded to NetSuite being our ERP and then all of the other fin apps, which we use as well around the piece, and everybody has to align with our global processes, procedures, and policies. It's it's just you you you can't operate unless unless you get to, bring everyone onto the same systems at the sort of level that we want to operate at. Sometimes it's not possible but luckily our size and the the kind of like the process we've made through the maturity of the business we've been able to do that. And we are, I think, two and a half years in now to the, finalization of NetSuite ERP implementation and the benefits you get from being on the same system and then also having everyone on FloQast and being able to have that visibility into all of the teams across borders, across jurisdictions, is is is just worth its weight in gold. Yeah. And I know we've spoken a lot about that before over the years, but just that visibility and structure and consistency is so key to everything. And then it's also about the process. So I think that, you know, the process and the structure behind that, I remember us talking about intercompany. And intercompany is always a a very global problem. But I remember you said it to me as well. I was like, well, you know, you just agree the balances. And then you figure it out afterwards, and then we fit the technology around it and we do the solutions. But a solid process of of how people agree with it also then enables that change and that consistency. Yeah. That was one of our major problems when I first joined Dept. Everybody hated intercompany. But, yeah, you bring everyone into the rules and say, right, end of working day three, everybody has to have agreed their intercompany, and it's forced through our consolidation reporting tool as well. They cannot submit their numbers unless they're below thresholds on intercompany reconciliations. You know, that makes it go a lot smoother. The rules and the structure as well as everything around is really tying everything together. I think what the stat... What the tools do is force you to kind of focus on the nonstandard and the outliers, and that's probably the benefit. Like you were saying as well, it's like you started in... Once the process is working properly, you start to focus your attention on more value added stuff. So what we're seeing is through implementing, FloQast is we've now got all our processes documented, and now we're moving on to the quality of our evidencing of our controls. And so really questioning that and really understanding is this appropriate. And, you know, we really struggled to get to that space before because actually the focus was mostly on getting the work done. I'm gonna take a break for our next poll question. And when we come back, we're gonna talk about build versus buy, one of the big debates within the AI space. Here's our next poll question. What's your biggest barrier to AI adoption? So one of the next things we wanted to talk about was buy versus build. And that's a debate we hear a lot of people having. What can I do myself? What do I need help with? And and kind of how do you strike that balance? Steve, I know you've been on this journey over the last number of years with lots of different solutions, and you mentioned some of them earlier. How do you make that decision on on what you buy and what you what you, build? Sure. I think the first question we really have to ask ourselves is what is the business problem we're trying to solve? We don't want to just throw technology at things. I think we have to ask, should we be doing things differently, and is there value in this solution? Now if technology is the answer, we typically buy where there's basic repeatable tasks, there's low differentiation, low integration, and there's a mature solution already available. So with our AI matching, FloQast already offered the technology and the platform. We actually did investigate with our Agentic AI team whether it made sense to do it internally, but the answer was pretty clear. There's already a product available. It's safe. Go with that. However, in the case of the agentic AI solution, we saw that we have proprietary data and processes. We saw that there wasn't a mature product available for us. We saw that there was high differentiation and we saw that there was high integration needed. So it made sense to build something ourselves. Yes. And I think that's really interesting when you kind of think through that when we were talking earlier as well as focusing on the right problems and focusing your teams right problems. If there's something that can be built and can be deployed pretty quickly, you're not taking your mental capacity of your team to try and take something and build it and reengineer something that exists. But I think it's a great example of something, like, really bespoke, really niche. That's where you need to think about the right things to do. I suppose, Nicola, what's your view on this spy versus bills? Yeah. You're absolutely right. It's... The speed to ROI, is is key to the decision making as well. And if, or we're looking at mainly optimizing the AI solutions already in the platforms that we've got and really driving value through what we're already paying for. And then we're also then considering, FloQast transform. We're we're doing some experimenting with the agents within within the system, but also alongside other options as well. So I think at the moment, because everything's moving so quickly LLMs is just phenomenal in the last six months, you know, to... I mean, if you look back a year compared to what they were, and and what they could do back then. So I think keeping options open right now is also one of our strategies, whilst making sure that we're experimenting with with all the options that we have available to us. Yeah. And I think I think that that pace of change is is is something that's really hard for people and especially with... Like, I I love, Steve, what you're saying is you come back to, like, what is the problem and are we solving the problem? And if we can define that we're solving a problem, great. That's a good point to invest in something. But if you're not really sure where people are using technology for the sake of technology, it's not really it's not really helping that. So I think that's a really good use use case because, yeah, the the speed at which things are developing, is there. But there's also the tricky balance of doing something because I'm also seeing teams fall into the trap of, like, you know, we wait for it to be perfect and we wait for the ideal solution, and then you realize you don't actually do anything in the first place. So a lot of people are like, oh, the data isn't here. I don't have this. I don't have this. I don't have this. But part of it as well, you just have to start somewhere and make some progress. Yeah. Absolutely. Just getting started, experimenting, seeing what the teams can, you know, develop and and build themselves. You also... You know, you've got that issue whereby if you're building it yourself, you then have to maintain it and keep it up to date. Whereas if you go with a partner and a solution which is already in the product, you know that they're maintaining it, and they will continue to develop it as well. So, yeah, there's a lot to consider. Yeah. I think as an agile business, we're very used to kind of testing and learning. So what we do tend to do is we focus a lot on discovery. I map out discovery for anything that we want to do in terms of our road map to make sure we have enough head space to really look at proof of concepts, really test it out, and feel is this right for us, get... Gather the information of the teams. We do showcases to then gather feedback from everyone and sort of listen to the different voices in the room to make sure that, you know, whatever... Whichever direction we choose to go is appropriate for us. But like Nikola, it's it's about testing and learning. Right? You're never gonna get it 100% right. You can't wait for the latest, you know, new thing to do. What you've got to do is try and start somewhere and then see how that evolves over time. So there is no right or wrong answer, and it really does depend on you. You just gotta be able to adapt with it as you go as well. Do you open it up as well to the teams so then, you know, here's an idea. We're gonna try this. And then people will find the errors, and then they go, oh, this doesn't quite work. And they go, okay. Well, we need to go away in this race. And I think that's really how change works. Yeah. We we focus a lot on retrospectives. So what we tend to do is gather everyone after we've done something and then listen to feedback. We use little post it notes and my voice and and gather all the feedback and then think about, okay, where to next? Because I think it is about that problem solving and really working through why something is a challenge. Because especially when you're trying to roll out new software or, like, a new process, you get resistance from the business naturally. So it's how do you address that resistance in a way that's inclusive and makes everyone feel collaborative as well? Big nod there, Steve. How do you tackle that resistance to change? Is it as much about the people as it is about the technology? Yeah. I guess we, try and tackle change through two approaches. We have, I guess, subject matter experts. We have transformation specialists in each of the centers who talk to the local business and stakeholders, really try and understand their needs, map the specifics, the exceptions, and try and work through them. And our general policy is global as possible and local as necessary, so we will be local if that is necessary. And the second thing is, I think as the ladies shared, right, we need to show the value in some of these products. So we're lucky. We have two regions which are really on board, very easy to work with. And the other region, understandably, because of the local complexities, is a little bit more resistant. So what we tend to do is show that it's worked in the other regions, demonstrate the use case, they see the value and then they quickly get on board. Yes. And I think there's a lot of that and especially, I suppose, Nicolas and a team like yours, I know even talking to your team, you've got team in The UK, you've got people in The Netherlands and kind of that mixed culture and and getting everybody to buy in is is is a challenge across that environment, is it? I I think it it can be, but I actually think in our team, we we have really good collaboration. Excuse me. And, yeah, everyone pulls together and everyone knows what the what the... You know, if we, articulate the strategy properly, then that normally means that you get better buy in. Yeah. So the teams will come together if they know why we're doing something, and they see the benefit from it. So I think it's mainly in the communication at the start and making sure you're very clear on what you're doing and why and the benefits from it, and then the team will just... They'll they'll come with you on the ride. Yeah. And it's it's interesting as well. Just this flip to the future because accounting, you can get very lost in looking at the past, and this is what happened yesterday. But it's taking people and kind of breaking that and get them to to look a bit further forward. Perfect. And and we're gonna take a pause here to take the next poll question. Our third poll question. How would you describe your organization's financial controls today? K. Now we're gonna get to the fun stuff. We've talked about the potential. We talked about the different use cases, but, we have to be boring and I can't just talk for a minute and talk about the compliance and and how we apply this. Because I think one of the things... And and, Nick, you probably know this is working in a creative agency is it's great that you can use AI to prompt a thought or to do something fun or to do something innovative. But when you're in an accounting space, there is data, there's security, there's control. Like, how how do you find that balance and kind of regulate how the team are using this technology? I I mean, it's also very relevant to the business teams as well because they shouldn't be putting any client information or data into any of these tools which aren't already approved by our legal team. So our legal team and compliance team manage all of this, and so we make sure that we adhere to their policies. So the group has a global AI acceptable use policy. It has approved tooling in there, which you are allowed to use enterprise wide, and then they also have a process for, requesting access to new tools. And depending on what sort of level of trial or if it's more of a permanent process, then they have certain ways that they will go through a committee where they review the tools, and they go through all of the, required review of the governance behind and security behind each of the tools before they then, let people have licenses. So, yeah, we have, training as well, which is being rolled out by the legal team. We've we've had training in our AI fluency program, which we've done within finance, on the risks on the... Of using AI and including hallucinations, bias, proprietary data use, etcetera. But the legal team also make sure that the wider business as well, including finance, is fully aware of all of the compliance requirements and risks. Yeah. And it is. And it's it's it's... We talked a little bit earlier about, like, explaining why. And I think one of the things as well is you also have to explain the risks and the reasons why because you can get a very enthusiastic member of staff saying, I can do this. This is brilliant. I can throw all this data into an NLM. And then you're like, woah. Woah. Woah. Woah. Woah. Woah. Woah. Like, where's where's the structure? Where's the detail? Are you doing all the right information got in there in the right order? So it's it's balancing that excitement and wanting people to experiment with also just, you know, the reality of how we work and how we operate. Yeah. And therefore having those guardrails around so that you can trust the outcome of these models, but also, making sure there's the critical mindset of the team. Because if you you you get an answer from AI, AI is always always very confident about its answer. And if you kind of prompt it to say, oh, have you thought thought about this or have you considered that? Maybe this has got an implication. It'll come back and say, you're totally right. Here's what we should have said. So it's getting them... The team to make sure they are always reviewing it for themselves. There's always a human in the loop, and and and they're making sure that the outputs are what they expect. Yeah. I think that question... And and it comes back to just professional skepticism and judgment because it's always... These models are trained to give you an answer and to to also look very helpful. But, yeah, I've had many conversations at mine as well where you can go and think a bit a bit like that or, I taught, my Claude a skill of just to think of it like a CFO. And I was like, if you give me an answer, double check it. And even if you double check it, then then check it again. But, like, even, you know, teaching some of those skills and some of that logic, I think it's really important that we educate people that this thing is trying to help you, but it's also you need to to double check the facts. I suppose, Diane, how how how do your team internally adopt and and manage the the regulatory... The regulation in this space? Very similar to Nicola, actually, because we have quite, established processes when it comes to us, particularly in procurement space. So if we are sort of buying something, making sure we have legal and procurement involved early on means we open sort of security, cyber, and, you know, AI compliance. And so that's kind of also matured over time compared to know, when we contract with FloQast before versus where we are today. Processes are changing all the time. And so it's making sure we involve the right individuals to make sure that from a security perspective, you know, we're happy to take on a certain toolset and there's just some, rewards in in addition to that. When it comes to that copilot, for example, yeah, I think it just varies based on sort of the different experiences with people and so how well they prompt the the agents and, what they get out of it and how they, you you know, critically evaluate it. We have done quite a bit of training internally as well to to fine tune sort of prompts and how we get the best outcomes from it. But it does come a lot with practice as well. It it does as well. And I think that prompting skill is really important. And I also think it's really important that you think about... You educate people on that and kind of the different ways to do it. Yeah. We've done, like, workshops. So we've gotten everyone together in a room and said, okay. Everyone do this task, and you do it in a different way. And you put in place some best practices, and you see how it evolves over time. So, it's been something we've been adopting over the last year, really, is making sure that we get together with the teams and have workshops to try and get the best outcomes. And that's more for smaller use cases to identify, like, the incremental wins that we could have. But then, obviously, when it comes to sort of, the buy, option, then that is kind of more controlled and a lot more sort of set guard rails in place. Yeah. Yeah. Interesting. We are... I think in our annual performance cycle, we have a question this year about AI fluency because, you know, how is everybody using it? How are you... And it's a good question for everybody to just sit down for a moment. I think, what am I using this for and how am I how am I impacting, how will it work? I suppose, Steve, how do you go through this in terms of this as well in this this multi complex jurisdiction? How do you think about the application? Because you you probably should think a little bit about the EUAI act... EUAI act. But also then, in such a global business, you've got lots of different viewpoints to take into consideration. Yeah. And especially in recruitment, right, where we deal with a lot of sensitive data. So I think my answer will be similar to the ladies in the... That without the right governance, we can't... Cannot scale the AI. I mean, there's no point having a big efficiency win, and then we have a data breach or audit finding on our revenue recognition, policy. So, yeah, important thing is to prepare the organization in terms of internal solutions and external solutions. So for external solutions, we'll have security questionnaires, privacy questionnaires. We'll also have our legal teams review. We have solution architects in place. And in terms of the agentic solution in OTC, there were several hurdles we discovered that we needed to jump. Of course, we're learning from some of the other departments who are a bit ahead of us on this, but it's a really interesting learning journey ourselves. So we definitely needed to, adhere to GDPR requirements, especially in terms of data retention and the right to be forgotten. So with this solution, we basically created a new environment. So we need to ensure that that was secure. So we worked very, very closely with our data protection team. We need to make the... Make... Ensure that the overall environment was secure, as well. So we needed to map where the AI was going to run, how it would interact and interface with our other systems, where access controls were, and then we needed to test the AI itself. Right? So we actually hired a a team of external consultants to embed hidden text, use small text in emails, and that was really, really useful. And with that, we managed to give the assurance to the wider business and the legal and data teams that we had done what was necessary. Yes. I think this is really interesting because it's also thinking like these are all new challenges as well. We talked about legal. We talked about procurement. Like this stuff is advancing and changing so quickly. Like, you need to build in time. So settings... As we kind of work through the conversation is you need to have the business problem that you're trying to solve, then you need to solve the business problem. You need to go through these new and probably different difficult conversations that you're having. It takes a bit longer because you need to prove the site and it's a different business use case. You need to get the team on board with it. And then, you know, just in in the accounting space, we need to think regulatory and all this. So it... It's quite a... Coming back to what we talked about at the start where somebody goes, oh, this is really quick. Just snap your fingers and off you go. You can you can deploy this technology. There is actually quite a journey to really get to driving value for the business, I suppose. What's that time scale maybe, Nicola, for you to like really go from thinking about something to really seeing true impact for the business? We do things at 100 miles an hour at that time. So maybe we're not comparable. Well, yes, but then we hope to not break the stuff either. But I don't know. I mean, it depends on the tool completely. It doesn't matter. So something quite simple. And I'm not saying FloQast is simple, but, actually, implementing FloQast was pretty quick for us, whereas, obviously, a bigger ERP is gonna take a lot long... Longer. And I think now with the extra requirements on the security, and the governance and and the data protection, You just have to make sure you make the time at the start because if you get that wrong and your legal team or procurement team say, actually, no. It's not acceptable for us. You've wasted an awful lot of time and energy. So, yeah, getting all of that and your ducks in a row from the start is important, and then you can go fast from there. Yeah. And I think that's it. It's it's usually... It used to be what's the business case and lots of value is gonna bring. But, you know, there's there's new hurdles you need to jump. But I suppose, Steve, you talked a little bit about things that you tried to fail. Is there any lessons that you'd have for for the audience in terms of, you know, how you move through that? Yeah. Absolutely. Probably similar to what Nicola was saying, I think we probably need to get different stakeholders on board earlier. I think everyone had their own agenda. We probably need to get everyone sitting around the table, understanding the what, the why, the how, to get their buy in probably earlier in the journey. It did take a little longer than expected. We probably found that the biggest challenge wasn't the technology. It was probably how we organized ourselves. Yeah. And as well, just with that, I suppose Diane just to to to take that on as well because we're seeing, you know, these elements as well. But, you know, we talked earlier with you about just the collaboration of getting all the stakeholders to buy in. Like, is is that really important that stakeholder management? It's not just about a good business case now. It's about the good business case, but then also the stakeholder management to navigate the organization to implement it. Yeah. 100%. I mean, you asked Nikola about the time scale it takes to implement. I think for us, it probably takes about nine months to twelve months to contract something. And so, actually, what it becomes is by the time you make sure you've gotten all your compliance and regulatory in line and actually contracted to get the tool, implementation is not always the the hardest part of that journey. But then it is getting people on board and making sure they're accepting of change. They feel part of that journey. They understand why we're doing what we're doing. But also just knowing that you're ready, because often we set about on a journey, and then you get there and it's like your week of implementation. You realize, actually, there's a couple of things we haven't thought about yet. And so it's taking those lessons, logging it, and really learning from it before you continue to scale. And that's what we are finding is every time we do something, it's like step... Stepping back and reflecting on what's gone well, what hasn't gone well, and why so that the next time we do it, we can improve on that. Yeah. I think that's really important in that and that kinda continuous learning. And and as well, you talk about in that time scale is it's also reminding people of the value because you could define and you could... You know, this is the value. This is what we're gonna get, and it's just keeping that momentum and reminding people of that because by the time somebody gets it, they're like, oh, I heard about this six months ago. That's like, yeah. But, you know, this is the time it takes to really get value from it. Yeah. Governance element when we started to implement different tools two years ago has changed. So that landscape's gotten, a lot more precise, and there's a little bit more clarity and transparency in terms of what is needed. So that's helped our case a lot. Okay. We're gonna go to our final poll question then. Then when we come back, we're gonna talk about what the panel is thinking about for 2027 and the future of their finance functions and also some closing advice for the audience. And our final question is, how far along are you in defining the future operating model and structure of your finance function? Okay. So to to finish off the conversation, I know we're not even finished 2026 yet, but I'm gonna annoy you and ask about 2027 already. But what are you guys thinking about in terms of the future of the finance function? Is it different skill sets? Is it different technologies? Is it kind of adaptation to change, I suppose? What what are you thinking about, in terms of kind of the future future proofing the finance team? Nicholas, let's start with you. Yeah. So through this year and last year, we've been getting a base level of AI AI knowledge across the team, and skills. And we are moving forward with our AI fluency program for next year, and we'll be work... Working on more advanced skills in AI. So really elevating the team skills and their abilities to to adopt these new tool... Tools available. But that also means that the the roles are evolving and the skill sets that the team need need to evolve as well. So yeah we're we're seeing people moving from kind of people who do transactional processing to people who are gonna be managing agents and oversight of outputs from agents and also looking at, design of architecture. So the skill sets in the finance team for us are going to change, and we want the team to evolve with that required change. So we're we're working on upscaling the team across the board to do that. So, yeah, that's kind of our main focus for 2027. And... Yeah. I think... Yeah. And I think I think it's really important because, like, we've talked about here, like, the technology is one thing, but then the people and the skills, the mindset to adopt it is there. And I... Yeah. Because the guy at the end of the K2, he works for a recruitment company if you need any new skills. So, Steve, you you can thank me for the plug later. But, Diane, I suppose, what are you thinking about going into next year? Yeah. I think so where we are in our journey at the moment, we've implemented full cost. We're implementing an accounting engine. Our focus is on data. So going into next year, our focus is on making sure that we have single source of truth in terms of all of our data so that it's a good foundation for implementing AI. And then second to that, his name is Nicola, is skill set and evolving the operating model of finance. Because as we're starting to transform our different functions, so what does that end goal look like? Because, actually, what you have is a mix of people moving between accounting and managing finance systems. And what does that look like for a day or so? It's... I was thinking about the finer details of exactly what it looks to do at your job in a transformed world and what skills that's needed. We also have a data transformation going on at the moment as well. So it does also mean bringing on additional skills that's like data ownership, data analyst skills, engineering within finance, and what does that need to look like. We know that we don't have all the answers yet, but what we're trying to do is make incremental changes along the way that will align to the end state vision. Yeah. Yeah. And I think that's really important, and we didn't really talk about it that much. I know there's other sessions in the program specifically on DaSytt, but I think that's... It's a really key point that you make because that's the foundation. And I think a lot of teams were figuring out, oh, like we talked about earlier, oh, we can put in this change, and then what we do is we push in the technology. We realize we never actually had the data to take advantage of this and people are kind of coming around to that. But I think it's also interesting what you said about the different skill sets and kind of the education of people because now it used to be, you know, debits and credits and deferred revenue versus accrued revenue. But now you have to have people figuring out what's an API call and who's this system connected to this system. There's a whole new level of, I suppose, technological information that I suppose... When I was... When we were starting to be accounted, you wouldn't think about it. But I think that's is that what you're saying? Kind of bringing that technology view into what the teams are thinking? Yeah. Absolutely. Because it's almost starting to merge between the data skill sets and the finance skill set, and you are having to need people in your finance team to understand data architecture and understand the landscape and what it means to be a data owner, what are the roles and responsibilities aligned to that. And you start to find, actually, as we start to get more information readily available at our fingertips, is that your business partnering type of roles become a bit more data analyst type of roles, and we're starting to think about what what should that look like. We're still quite a couple of years away from that, but we're starting to think about how do we lay the foundations now so that the skills are slowly starts to evolve rather than being a big bang approach and and having to transform skills massively. And it's the speed of information that people get now that they need to adapt to too, as well as that invisibility of how the data is kind of coming about underneath the the layers as well that they need to understand that, but then also be able to trust the outcomes as well having done the understanding of the basics of how it's working. Actually something I learned about recently was AI burnout. Right? So as the AI automates a lot of the basic tasks, right, people are now left with those high value added tasks which are the more difficult ones And if you're doing that all day every day when there's there's a risk of burnouts, we also need to manage that with our teams. It's a live conversation. We don't have the answers at the moment. I guess we learn by doing. It's it's interesting, listening to Dan and and Nicola in terms of their journeys in terms of the the teams. Obviously, we're a recruitment company, and we also need to advise our clients on what the future shape of our finance teams and other teams will look like. And we're in a privileged position in that we're actually doing the work, and now we're having some really interesting conversations with the business to prepare our recruitment consultants for that moment when someone says, well, do I need this role or do I need something different? So we're seeing that. So in terms of our priorities for next year, a large part of it is is shaping the team. We're gonna roll out the next steps of that AgenTic AI solution. We're gonna be using FloQast AI matching. And what's very, very interesting is that there's going to be a new version of Oracle NetSuite, which is called NetSuite Next. I think you're probably having the same conversations, which is going to significantly leverage AI, and we're looking at it as an opportunity to reengineer our finance processes. So take a step back and think, well, what do we really need from finance, and what's the most efficient way, and controlled way of getting there? And we need to shape our our teams around this new technology. Right? So that's what Adi said. Right? We need people who understand data. We need good accountants. Right? They need to be able to challenge the the outputs. Right? They also need to understand the business. They need to be really good communicators, right, to transfer that knowledge, to the business as well. So, yeah, we don't have all the answers, but we're in plenty of workshops at the moment discussing these items. Interesting. And anything to add, Datta? I think the challenge as well that we all need to kind of get to grips with is where you talk about we need more senior people, but where are these senior people coming from over the next few years or so if you're not... If you haven't got the triangle anymore on the pyramid, of the of the workforce. So you've not got people like you know learning on the job like we all did in in our early careers and building that knowledge through and you're expecting them to come in at a senior level. I think education's got to change you know the the accounting firms and how they train their people have got to change so that people can come in with those more senior skill sets that we're expecting people to have to interpret the AI and to manage the systems and be technologists. So I think that's something that is going to have to be got to grips with over the next few years as well. Yeah. I I completely agree with that because I think the expectation is going to change and what people are looking for in the output is there, but that's still that critical thinking and decision making and pushback is really important because I think one of the the uncomfortable things about being an accountant is sometimes you have to tell a message that somebody doesn't want to hear and deliver that truth to them. So I think that that position that will still remain important, especially in the world of AI is giving you all the answers you want at the end of it. I suppose just to round out the conversations, it's been great. We've covered a lot of grain. I suppose what will be to just come down to a work day in the panel, but I suppose what will be one bit of closing advice you'd have for the audience? I think we've talked about data and getting your data right and I think at that we feel pretty comfortable that we've got a good level of data quality and we've been working on that for quite a few years, so you've got to have that right to start with. And then I think it's about just getting started and empowering your team, so you know letting those team members who've got that pioneering attitude and they want to experiment, giving them the tools to do that and giving them freedom to do that within the guardrails and the risk profile that we allow, But just getting started, and seeing what they can achieve and seeing what they can then disseminate across the team. Brilliant. Thea? Yeah. I absolutely agree with Nicola. I think it's starting small and then iterating and sort of growing with that because it's hard to keep up. And so starting small allows you to then build off it and listen to the team and the feedback. User experience is important because while you might, you know, have a demonstration, it looks all great, thinking about how it's relevant to your organization and what that's going to look and feel like is quite important as well. Yeah. Build that success, build that momentum. And, Steve, take us out. I'm gonna sound repetitive, but, yes, it's that first step forward. I think what's also key is that take this opportunity, if you can, to look at standardizing your processes. Processes. We've we've learned the hard way, or we had a lot of process differences. So... And in terms of that first step forward, what does it look like? I think it's identifying an opportunity where there's real value and you can measure it. Right? So, yes, of course, we have to get all the fundamentals in place, but it's important to to take that first step. That's why I think it's important you kind of evolve that role as well because when roles start to change, you need to start to think about where where does it move and how does it evolve through that transformation as well. Brilliant. Okay. Well, thank you very much, Nicola, Diane, Steve. I I hope everybody's enjoyed the conversation. I think we've got a lot... We've covered a lot of topics there from AI. We've learned about buy versus build. We've learned about what real change and lasting success looks like and how these guys have taken these examples and turned them into real valuable outcomes from their business. Thank you so much, and enjoy the rest of the day. Up next, a quick break, and then stay with us for a session on how account