AI Readiness in Finance: Hype, Hope, or Here?

On the topic of turning insights into action, that's where we're headed next. Our next session, AI readiness in Procurement Hype, Hope or Here. And it takes that vision that we just talked about and brings it to the front lines. Joining me on stage here a moment are finance and IT leaders who are rolling up their sleeves, figuring out what's working, what's not, and how to build the right guardrails for AI adoption. So as the speakers are coming in here shortly, I'm going to get started with a few introductions so you know who to expect. Alex, lovely to see you. I'll begin with Alex. Alex Costa is the Senior vice president of IT and security at PSC Group. Alex is leading PSC's digital transformation across the US and Canada. Next that came in the door there is Damon Norris. Welcome, Damon. Damon is the Vice President of Finance at Basis ed. With over 25 years in finance and operations experience. Damon leads financial Strategy for a $250 million network of top performing schools. And I'm thrilled to see Elizabeth Parrish come in the door as well. Welcome, Elizabeth. Elizabeth is the Director of the Enterprise Vendor Management Office at questrade. Elizabeth leads questrade's partnership with its vendors to drive performance innovation and business value. Welcome, everybody. Thank you for joining me today. Thanks for having us. Thanks for having us. Good morning to you. Good morning. I hope you got to tune into some of our keynote there earlier. Some really inspiring things to really get our brains warmed up for this conversation. So I think it's the perfect, perfect preview to what we're going to chat about. Well, to get our conversation started here, let's get a little bit of situation awareness for our organization and for everybody online today. We've discussed already that AI is a journey and that organizations, even teams within organizations, are each on their own path of adoption. So to kick us off, Alex, I'm going to come to you first. Can you tell us a little bit about the catalyst that began PSC groups and AI transformation? So for us, it really started with, I mean, being in it. Right. We are adopting technologies that have AI embedded for some, some long time now. I mean, it's not new for us in terms of how we do security. Right. And all started because of limitations that we have in terms of personnel. Right. So in reality we, we have a limited amount of resources in it. So the tools that we incorporated in our ecosystem were related to how we review emails coming in or how we do data loss prevention and how we review our infrastructure and server monitoring. We utilize AI tools that learn Behavior and learn how many things in relates to how we manage our environment to help us, you know, augmented team. That's pretty much how we started right from there. Of course we start adding other tools and agents to address specific use cases. And the vision was for the most part to identify okay, instead of just adopting AI and trying to make AI fit use cases. The approach is different. Right. We identify use case and say, okay, how can we leverage AI to address a specific need or the specific gap? So that's how we started. We have a few things going on in the finance front and then the business development front where we're using agentic AI to address some of those use cases. But it always started within it and doing a lot of the back office kind of work that we needed to get accomplished with the limited resources that we have. Amazing. And Damon, can you share a little bit about your journey as well? So it feels like Alex's journey might be in the rounds of, you know, phase one to phase two of that transformation framework we just saw. And probably I know Alex is super ambitious to get into that frontier. Phase three. Can you share a little bit about at Basis Ed, an education institute, you know, what was the catalyst for your organization? Yeah, I think, I think being in education is a unique area. As many of the people on this will understand. The funding is fine, finite, right. I can't turn to, to another product and raise my prices to get more revenue for us in order to, to do something. And so our money is very tight when it comes to the budgetary side of things. And so it was about how do we do things with the same amount of people, but expanding the network and expanding our reach to bring our education and our schools to, to all different areas of the country and expand that, that education for the kids. But you have to do it with the same amount of money. And so it's how do you get the same headcount of people in a department to be able to handle what you probably need 20 people for? And so we started looking at different, different ways and I think the keynote speaker was, was spot on of trying to figure out how to get teams to manage agents and use the AI to help you do some of your work strategies. Work smarter, not hard, work harder, not smarter, harder, smarter. I screwed that up, I apologize. But, but, but, you know, trying to, to really optimize our teams so that they can get more work done and do it very smartly so they're not touching every invoice. For example, if it can flow through, it can flow through, but the human interaction still has to be there because I think AI hasn't gotten 100% to where we needed to be for that. But we're, we're going on our journey to that and it's exciting to see. Yeah, thank you for that. And you're right, we do have a lot of education, you know, customers and, and community members online today. So please do share what's resonating with, with you in the chat there. We'd love to hear. So Alex, from that manufacturing space, we've got Damon there from education and now we will turn to hear from Elizabeth. Elizabeth, I know questrade is on certainly an AI transformation journey. Can you share a little bit about maybe when it began and a little bit about the catalyst that began it? I would say right from the beginning. I can't speak for the developers and the engineers at questrade and I wouldn't want to, but I know that very early adoption with a lot of the tools that have become available, especially any tools that are going to again, not replace a human body, but have an extension of them. So when it comes to, you know, repetitive tasks, anything that we can choose to automate. So we've been trying to leverage where we can, various tools that provide or are maybe, you know, piloting some sort of new AI type tool that can create efficiencies for us. So we've sort of grasped onto a lot of those at the same time while continuing to develop internal policies. And then when it comes to procurement, because I can speak more of that sheer volume. Right. Like, so my team wants to leverage anything AI related to assist us with volume because to Damon's earlier point, we don't have a thousand people that are working on this every day. So anywhere that we can leverage something to automate or give us, going to use the pun here, insights into something, we prefer, you know, to be able to go and find that information on our own without having to go and ask someone to sort of run a report or export this data and then put together, you know, a pretty little pie picture for us. So we've been looking for anywhere basically that we can create efficiencies and be an expansion of someone's role, not a replacement of someone's role. I love that. And you touched on it, you know, around the usage of what you're leveraging in Procurify around AI. Can you expand a little further and tell us about your team's use of the curify spend insights now everybody online would have seen a little glimpse of that in the previous session and of course they've seen it on our website and through our sales and customer teams. But can you share a little bit about how your vendor management team in particular leveraging spend Insights and any use case do you think be relevant for the audience here? So I use this a lot and the procurement team uses Insights. I'm going to park sort of AP for a second because it's more so around the procurement, but in finding trends, utilizing information that's now readily available to me, as soon as I click on that screen, it'll give me an overview. We also have multiple entities, so from my perspective, I like to see what entity is spending and I found that with Insights, I just click a button and it's all there and it'll show me. And I think it refreshes daily or multiple times a day. But it is real time right now that we're going to get in regards to open purchase order requests, things like that. The other thing I know I utilize it for, and I believe that most of the team also looks at is when it comes to the approval, like time to approve. And I say that because from the management perspective, who do we need to go knocking on their door to tell them that the approvals every week or every day sit with this team for a longer amount of time and not to sort of get on their case about it? It can be twofold, where we're also asking is there any other information that we can provide you to expedite these types of approvals. But having the ability to see the time to approve and be able to just have that right there for us, I think also gives us an ED on our side to be able to go and talk with those teams to see if we can expedite anything for them or help them with information. So from the procurement perspective, I use and my team uses it a lot to view trends from an AP perspective, I'm just going to be honest, the team is really early adopters in using Insights from an AP perspective. The AP team and procurement teams are also looking forward to the next phase when we can pull out the data and automate the creation of purchase order requests. That's endgame ideal for myself and for my team. Yeah, and you mentioned the team there. How has the change been adopted by the team? What was the team's, you know, what was your process with your team to drive them towards leveraging spend Insights to then drive those efficiencies on time savings as you were identifying for other teams? I would say that some folks need to see something be done by another in order to see how it works. I think a lot of us live life that way. Right. Well, I want to see you do it first and then, and then I'm going to do it. I will say that we had a full blown demo from Procurify. I did ask a couple of team members to come in and sort of spoon feed us and the rest of the team about what we could leverage it for and what the purposes that we could use it for. And I will say there are probably a couple of team members that are using it more than others because they're waiting for sort of, you know, feedback from other. Hey, I just went into Insights and I was able to see this. I used it last week when we actually had Procurify on site so that I could say to the team, you know, I was looking for a certain purchase order that I couldn't find digging because I didn't remember which month we created it in. So that's something that I use the analyst like my side analyst in Insights and just ask them, you know, can you find a purchase order request that went out, you know, for this amount, for this type of spend under this entity? And it'll give me a bunch of different ones. Are these what you're looking for? But between it, I was easily able to find what it was that I was looking for as opposed to searching through months and months or years of data, not even remembering sort of okay, well what was it purchase under what category at that time? Because also, you know, there's changes that happen quite frequently. Yeah, I mean that's incredible. And if you could summarize, how did you quantify the impact that spend Insights has had for your team? It saves time. I don't think that there's any better way to describe it except for time saving. And also, you know that you're getting that information if somebody was manually pulling that data. There's also room for error. So I would say that, you know, saves time and, and not that you don't have to review the data, but you can be in, you can be assured that it's accurate data if it's in Procurify and not maybe a number that's come out of Procurify extracted, gone into another spreadsheet and then maybe there could be a human error there as well. So those are two of the biggest benefits that, that the team and I see. Amazing, Amazing. Well, I'll just let the audience know. Take a look in the Chat. Our colleague Michelle will be able to share some great content. If you want to do some further reading on Spend Insights, she'll be able to share some details in there. I'm going to move to Damon. Tell me a little bit about your team and what your early trials in AI have been. I know you've been leaning in on the AP side. Could you tell the audience here, I know it's quite the journey, so maybe you'll have to summarize it a little bit, but a little bit about the pains that you experienced, what the solution is that you're testing out and a little bit about how that journey's gone so far. Yeah, I think on our end it's a two pronged thing and I think Elizabeth has kind of hit it in a couple of ways early in our last answer. But it's a lot of trust. It's a lot of trust for the people in the system and knowing that the system is going to do what it's doing. We have at basis we're, we're a bunch of high performing people that want to make sure everything is right and there's no errors. And with that comes control freaks. We're all control freaks. And so getting people to trust the AI is always the key. And I think for AP in general, and we're kind of a little bit unique because our AP and procurement department are pretty much the same department led by one awesome manager. But getting into Procurify and starting to use the AI for vendor recognition and we're still not using 100% of the power that Procurify or AI even offers us because we're still building that trust. But I think every day is a step forward as we look to really turn on the AI power and as Procurify continues to update the AI. And I know one of the things in the roadmap, and I think you guys may have talked about it earlier, was automated POS that would get placed for us in the future for invoices that come in that didn't need an invoice and things like that for the AI and getting the trust built so when those functions come around we'd be able to hit the road running. But the AP department's embraced it very well. They've. We've come from another software which was a. I won't say the name but complete and utter nightmare. We made it work versus having a system that worked for us. And I think that's the key is right now we're using the AI. We're using Procurify. We continue to get better and better with using it and getting things in there and trusting the system and trusting what it's doing. Because again, like Elizabeth said, it's all about human error. The more that I have of people touching an invoice or matching invoices, the more we see, see issues. But we do have a complex chart of accounts. Not to lie. It's amazing how complex it is. And so that creates some challenges with making sure the system understands what we want it to do and how we wanted to do it. So it's been a journey of shaking people and saying it's okay to let go, trust the system and trust AI and getting them there. So I think we're, we're in a much better place than we were say 12 weeks ago. I think for that, which is definitely a good thing. You mentioned a lot about trust there and we were absolutely going to dive into how do we trust AI? What are the guardrails we put up? So we're going to get into a little bit of that. Meet now in a moment, but something that I was wondering, so you've mentioned that journey you're a few months in. Have there been any lessons that have really helped your rollout? Anything that as you say, you've got some cautious team members we're dipping our toes in, probably even watching and seeing a little bit. What has really enabled your team. Was there anything that you would advise other teams listening that are dipping their toe in and maybe have that cautious community around them that has worked or anything that you've led that you feel like had a positive impact? Yeah, I think, I think one is for their boss, who is me, to step back and listen and not just say trust the system, but listen to what they're saying because they're the experts in their role. Right. And so listen to them. It's a management thing, you know, like it's, it's not, it shouldn't be a top down force of going into AI as much as it should be an encouragement to adopt it and adopt the AI and the system in a safe manner instead of sending your financials out to, you know, some AI bot that's out there that can give them to everybody. But I think, I think the first part is listen, but also, you know, ask around. We worked with Procurify in setting things up and so one of the questions was, you know, who else is like us? Can you ask them how they did this or how they looked at it? And so utilizing, you know your, your peers that are on procurefy or using AI, because it doesn't necessarily have to be procurefi. I know that's what this, this session's about. But, but listening to others and talking to others. I know my chat blew up just a second ago with Elizabeth talking and saying we need to meet with her and ask her how she's doing. She has multiple entities. So, so really reaching out and partnering with people like Alex and Elizabeth and saying we're not the same industry but we're doing things that are kind of similar. So how do you handle this? Or you're a little bit further in this journey. So how'd you get here? And I think that's the key is asking questions. And then the biggest thing that I've taught my team is it's okay to fail. And as long as you fail and you learn from your fail, but fail quickly, pick yourself up and learn from it. And we've done that mistakes in setting things up, we've made mistakes in some of the things that we've done. We've quickly learned from them and continue on. So I think that's the biggest advice I could give is when adopting any of this stuff is to listen to everybody. Everybody's opinion should count because they are going to see something. But don't be afraid to fall, don't be afraid to fail. It's going to happen. There's no perfect row out in any shape of the word being perfect. I mean, everything's going to have its hiccups unfortunately. And I love that the spend culture community is alive and kicking there. You mentioned your team. I know, are online and they're already seeing opportunity to connect with Elizabeth around what her team are doing. So I love that. That's what it's all about here at Pulse. Right. It's really building that knowledge share between everybody in this community so that you can all proceed together. So keep that going. We're going to get into the topic of compliance because that's a huge area that we're thinking about about. And Alex, I'm going to come over to you as the leader in IT at PSC Group. What kind of systems or guardrails have yours has your company adopted on your AI journey? And we'll talk a little bit about whether it's accelerated or, you know, hindered innovation in any way, but tell us a little bit about what you've set up to help protect your organization. Yeah, so I'll put my security officer hat real quick. I mean, I think the Main challenge, right, is with the, with the advent of AI and all the different agents available is the leak of information. Right. So I'm pretty sure we're going to find examples everywhere where somebody wanted to upload a contract to ChatGPT and do a review. Right. And with some information that not necessarily should be there in a public space and depending on the model that you're using or the services that you're using, right. That information can and cannot be use to train the model. So there's a lot of concerns that come with that. So we, we had to implement a few DLP security things that we push down to every single PC to track utilization of AI and to be able to make sure that information is not just flowing and. But it's the same reality with email, right? People share stuff that shouldn't be email. It's. It's just another catalyst or another path for information to leak and from our organization. So definitely in terms of how we implemented guardrails, I mean there's a ton of technology available that came out with the advent of AI to help secure the environment. We developed our own AI internally, all agents for a few different areas within the business. We have one for business development as an example that is really is leveraging the ChatGPT5 model in the back, but we kind of create all the graduate required so the data doesn't flow outside. Is our own agent to leverage in that model. It's really used for RFP responses. Right. So it's a model that we train specifically in how PSC does businesses and, and everything that we do. And it's really to help assist the business development team respond to these questions quicker. Right. So the expectation there is not that it's going to do the whole thing, but maybe get them 60, 70, 80% there when responding to those RFPs. But again on those cases, I mean we just had to engage and see what the proper architecture is on how we implement those models to make sure that they information secure. Right. It's just not flowing everywhere. And specifically for the BD team, we talk about our rate sheets and everything that we do when we're bidding for a project. Right? So it's really information that can leak. So and then, and finally, I think as anything in cyber security, you have to train your users, right? There's no way around is the weakest link, right? Because if people really want to leak information, they'll find a way. Usually people does find a way to go around systems and security and then you have to train them and Give them the proper tools and the right procedure and explain to them the risks. Right. So we took different fronts. I mean, of course we implemented the technology that's required to maintain things secure, but we also pushing different training and education to our folks to make sure that they're using the technology properly. I love that. I absolutely love that. And Damon and Elizabeth, I'm going to come to you both next. Can you share a little bit about either what's happening in your organization or in your community in the education space, what you're learning or the insurance space, what you're learning. I'd love to hear about any guardrails that you're thinking about, at least based on, to Alex's point, how to keep, you know, confidential information safe while still allowing for that motion of agility. So I can say on our side, every day, if not twice a day, there is a vendor, whether it's, you know, a SaaS, a tool, something where they've come and they have a new AI bot or they have some sort of capability that, you know, one of the many, many teams that questrade will want to utilize, of course, and we don't want to prevent teams from utilizing. But to Alex's point, there's a lot of due diligence that actually goes on behind the scenes before we'll even allow a team to proceed with like a pilot. When it comes to data, like Alex is speaking to, like DLP is huge as well as it has to go through. There's, there's a separate. And we're still, you know, as, as AI evolves, things are changing constant. There are a key group of, you know, SMEs and individuals who review what that tool is going to do. Where is it going to take our data and train it for models in another way? Because that's obviously something that we're not going to agree to. But where I've also seen that there has to be not only a due diligence review, but from a technology standpoint, and I'm not the most technical person in the room, but if it's some sort of legacy platform that then, you know, there may be an API that our, our technology team has like built a wrapper around or customized in some sort of way. And then you want to turn on this AI type of tool, it may not work, it could actually impact your system, it could cause it to crash. So I think that there is a real level of not only due diligence when it comes to like, governance, oversight with data. And where is it going and who's seeing this and who's using it, but on the other side as well, is it going to work? Right. Like, is the technology, the way that the infrastructure may be built today, going to work just by flicking a switch, or is there something else that you maybe need to build or integrate in order to be able to leverage it and actually get the right usage out of it? So we have a lot of that that goes in on our side, which obviously can take some time. Maybe not as much time as the people with eager fingers willing to wait, but. But we have to, like, it's part of our job. We have to do our due diligence on our side to ensure that it's a good fit before we're just going to allow anybody to turn something on and start playing with it. Also because of as how new it is and how it is ever evolving and that we don't sort of know where this ends. Right. Mm. And Damon, what's your experience in the education sector? Yeah, I'll ditto basically everything both Alex and Elizabeth said. We have. We have an amazing IT department that locks everything down, which could be frustrating at times. When you're in the finance department, you're trying to move fast and you're trying to adapt to the newest and latest things that make your life easier. And they say, hey, wait a second, we can't do this. However, education has been the victim of multiple cyber attacks and stealing of student data, as well as parents data, because we have so much data and so we have to be very careful at what we do. And, you know, we. Cyber security is huge to us and making sure that everything doesn't allow them into our systems and allow them into our network. Because not only does the data have to be protected, but we also have to protect what the kids see and what the kids are exposed to as well. Or the students, I should say. They're all kids to me. But. But seeing them and going through the processes with all that security, I think the biggest thing on our side is the guardrail of. And Elizabeth said this. I get 50 emails a day with, hey, we've got the latest and greatest tool for you and education. Like, everybody seems to think we have tons of money. I don't know why it is, but everybody tries to sell to us and they always have AI. And the question then becomes, is it truly AI or is it something you're just branding to be AI and does it actually help us actually do something? And then it comes down to we have a lot of intertwined Systems that have a lot of customs integrations. And so inserting something to this, like Elizabeth said, can break the entire system. And so we do a lot of reviews, which I'm the eager guy that's like, hey, let's just launch this. And I've got the IT guys help roll me so that I don't do anything foolish. But we definitely take security very seriously. Even with Procurify, it was vetted out pretty significantly because of the data that we're putting into it and looking at everything and making sure it's insured and all of our data is insured and so forth. So we try to wrap our arms around everything that we do with our data to protect it as much as we possibly can. I love that. And I'm just thinking a little theme that's been coming up here is the idea of how do you balance that pressure for speed and automation with security and compliance? And I love the spread of speakers we have here. So we've got, Alex, we're going to go to you for your IT and security hat, and then we're going to go to David. Elizabeth, as customers of these systems and really we want to get to know and for everybody here in the community, how do you balance that? How do you come together as a union and find that happy medium? So I'd love to get your perspective, Alex, to kick us off. Well, I really think that if you have the proper framework in place in terms of guardrails, right. It actually moves faster because instead of slowing down, because I mean, the trust is kind of built in into those guardrails. Right. So these are the things, the expectations that you have for the tool and those answers or those questions get answered initially or throughout the process. Right. In the beginning of the process. Right. Does it fit exactly what your security guardrails are? So if there are multiple tools being evaluated, right. That you start taking out the ones that maybe don't fit, or maybe they're not enough, they have enough maturity for your organization. And then you kind of start focusing on the ones that can provide real value, but at the same time they are aligned with what you have from a strategy standpoint of how you are securing things, I think balancing speed and governance, I mean, you can pilot things quickly and if it goes through compliancy and shows real balance, shows real efficiency, you can scale. And then you also define what the acceptable risks are. There's some tools that might have some very low data sensitivity where you can pretty much take more risk. And there are others that can be at higher risk for organization. And then for those, the due diligence may be different. So it's, I think it really, you know, there is, I mean, I'm the IT guy here in the room. Right. So you always get that the feedback of like it is trying to block us. We want to move forward, we're going to do our own thing. And in reality, I try to partner with the teams that we have here, and I think at PSC we do. Our leadership has the approach to CIT as a business partner. So when I brought in to those conversations early, we can really help expedite the process for other teams when they're looking at technology that they want to implement. And that goes to everything, not necessarily just AI. I love that. So a real balanced view is that we're really actually trying to help you and the rest of the organization to move quicker. And I love that you were talking about accepting acceptable risks depending on the area, so you can provide that acceleration path where possible. Elizabeth, what's your perspective on your side? I know, you know, it's probably more on the cautious side than the, the accelerated path view, but let us know what's happening. So when you said, how do you balance it? I was going to say with a smile on my face, otherwise, people are coming for me. But I really want to touch on what Alex said because I absolutely think that we've also started to take on some of that even as procurement, where if we like, we've done this a bunch of times now. So we do also know a lot of the questions that are going to get asked by like ITGRC or whoever it is. So even when we're looking at a demo, we'll start to ask some of those questions about especially like, well, what do you do with the data? Or how does this work? Or is there another integration or another layer and we'll get to cost maybe separately, but. But yeah, no knowing what questions to ask ahead of time. And the other thing that I also agree with Alex on is the sooner the better, engage us right away because we're the middle people in between technology. Right. And whoever's requesting. So for me to go to the technology teams or whoever it is as part of like the AI team or to speak to ITGRC or whoever it is, the earlier we know that, the more that, the faster that we can at least get those conversations going and get those reviews started. What I find is, you know, your requesters will feel more irritated maybe with you if there's a delay and I find that that happens when someone comes on a Monday and says, oh, hey, I've been talking to this vendor and they've got this great tool and we want to turn it on. So are you good with that? It's free and it's like, whoa, whoa, whoa, whoa, whoa. Have you been talking to them for two weeks? Because then engage us earlier so that at least we can, you know, surface the right SMEs to the to to be able to speak to this and review in a timely manner. And then that way, you know, you're satisfying everyone who needs to make sure that they do their due diligence and have oversight into what we're doing. And then also the requesters that want to start utilizing AI because it's going to create some sort of efficiency or it's going to make something easier for them. So I also look at that as money lost every day that we're not doing it where, you know, where we could. So it's tricky right now to find that balance. But I agree, like the sooner that we get engaged about it and knowing a lot of the questions to ask. Early onset. Yeah. Very sound advice for audience online. You've got to engage your stakeholder community right from the outset so that they can work with you and then not perceive to be a blocker as you're going through the process and really eager and start to lose that momentum. So really smart advice. And Damon, what's your experience with this? I think, and I echo exactly what they've said, to be honest with you, so I'll try to keep it short, but I think that, and Alex said, you know, it kind of gets the bad rap for slowing stuff down. Well, procurement also does. And now we're standing shoulder and shoulder with it. And like Elizabeth said, we know the questions before they're even asked because we've worked on these things before. Basis ed, for example, multiple entities, 40 something schools across the nation. And when one school comes to us and says we want to do this, it's probably not the first time that we've been asked it. And so we know how to get to the stakeholders and get the right people in the room and ask some of the questions prior to going to them so that I can go to, you know, my version of Alex here at Basis Ed and say this is what they want and this is why, and here's all the information. And then they still may have some follow up questions, of course, but at least we're speed. We're now a speedboat versus a Slow boat on the ocean trying to turn. And I think that's the, I think that's the big thing is really being agile to everything and understanding that, you know, in my case, the schools have a need. They're desperate to get this need for whatever reason. And if, if we delay that process any further and not communicate to them, they're not going to trust the KIP process, but they're going to go do it on their own. And then it becomes a bigger problem for all of us. And so communication is the big key for us on our side in the education, at least in my education world, I should say. I love that. And we're going to move on to talk a little bit about resourcing now as a theme. You know, we talked a little bit about it there a minute ago, Elizabeth. But with AI evolving so quickly, how do you evaluate new vendors or tools before scaling? So one part, step one, if you were to give advice, is to bring your stakeholder community in early. But what criteria or guardrails are you looking at in your vendor management office? From a procurement perspective, all of these tools that are coming out right now that are free to you to pilot and use will likely come back and be given to you next round or upon renewal at a cost. That's something that my team has been noticing for about a year now. There were a lot of, especially early onset of, hey, we have this and you know, we've done our diligence and teams are using it. Those teams then will become dependent on it. And why wouldn't they? Right? Because again, it's saving us time or creating efficiencies, but something that, you know, not even out of. And I'm kind of stepping forward here a little bit, not at evaluating the tool, but like, because just from where I stand on a procurement perspective, we're looking at spend, right? And so when it comes to the teams utilizing some sort of tool that has been maybe given to them as a pilot or as early adopters to test out, and then they become dependent on it. Now, then when we have to purchase or spend money on that tool, that's when there will be a lot more questions that will surface in regards to, okay, well, what's the return of the investment on this? Right? What is it that you're achieving by utilizing this tool where you're just playing with it maybe at first, and so you'll be able to come up with all of that reasoning behind it, but that also can prevent time to market as well, like to turn something on maybe if it was just in a pilot form, because then, you know, we're going to actually put it in a contract, we're going to pay for it, and then people are going to need to speak to the value that they're, that they're gaining from utilizing that tool. So to all my procurement friends out there analyzing spend, I highly encourage you to look out for these, because I'm living and breathing it for the past six to eight months. Yeah. And actually, I'm going to deviate from the next resource question in a moment because you touched on something around quantifying ROI from AI. So I love to hear that. So, as you say, you find these tools, you get really, you know, entrenched in. You love them, it's really valuable. But now you're also thinking about, okay, like, what's the new cost for 2026, 2027 and beyond? How do you quantify ROI from AI? I think certainly team efficiency, is it redeployment of, of staff that can now do other jobs in addition, you know, thinking back, it's. It's exactly that, Brona. It's, you know, what else does that free up the time for doing, like, redundant, mundane tasks, to be able to do other things and leveraging AI in that way? So that value alone, you know, how can we even sort of count that? But I also would say that when the teams are evaluating sort of what the, you know, what the value is that they're getting for that spend, I heard, I think, from the keynote speaker, and I find that that exists here too. We also don't want any resources to think that this is going to replace their job, so it should be an extension of them. And that's why when the teams are working on what is the value? A lot of it is, well, I can have this running and doing this while I do this other thing. So you're not replacing a resource, but maybe you don't need to add five new resources that year. Yes, it's about moving smarter, more strategic and being more strategic with your resources. And Damon, I know you've got a perspective on this around. How are you reskilling or reorganizing maybe your finance team to work alongside roi? ROI with AI more like really juice that roi. But then also, what are some opportunities that creates for you within your team structure and your organization around the breadth of work that you can achieve? Yeah, so. So we're looking at a lot of different things, right? So as we look at AP and, and as we adopt more and more of the Automation process and the, and the AI that comes with procurify that can, you know, simplify that track. It's going to free up hours of time for humans or capital of our, of our staff. And so how do we redeploy that? And Elizabeth said it straight. It's like how do you, we're not trying to eliminate you, we actually want to keep you and potentially either promote you out of an entry level job potentially here doing just processing invoices or to you know, something different, retrain you and have something better for you, or let's use this as an add on a bolt on to them. And I had somebody actually explain it to me of being basically being able to have that robot sitting next to you helping you do the entry level things while you're still processing and doing the stuff that needs a human to look at it or to interpret it. And in education at least I would assume this is everywhere because been in manufacturing before too. But when you look at that stuff we have very complex structure as well as we're dealing with a vendor that does like special education services across all of our different schools and different states and different entities. And it's really hard to train a computer on how to break that down and where to put things. So you still need some human interaction. But we're working as we can to now transition our AP people, analysts as they can onto procurement and teach them to look at spend rather than just paying the bill. Now doing exactly what Elizabeth was talking about. How do we control our spend? How do we make sure that we're getting the best deals that we can out there using the right contracts and, and doing the right procurement steps and, and really transition to help the help. This will save money rather than we're just paying bills and, and being a processor. So it's really looking at the human capital and how do we redeploy them in a proper way. And in a lot of cases that also comes with a little bit of additional cost too because you're moving them into bigger roles that are more responsible, higher roles, whatever you want to call them, and that comes with higher pay. And so you're trying to balance that ROI with potential payroll impact as well. And so we're seeing positive results out of that when we do that. Lovely. I love that. And you know, just, I don't know where the time has gone but we definitely are going to move into our final section of our conversation today and really about that future state. So this is really looking at the, the glass ball or whatever way you want to think about it is to what do you predict to be coming next or what are you anticipating and hopeful for? So over the next, let's say, gosh, a year goes very fast. So let's say a year or two years which AI use case. And Alex, I'm going to begin with you actually on this one. You know from your finance team that you're obviously partnering with which AI use case in finance or the procurement team of that will deliver real roi. Well, we, so as we continue to grow and Elizabeth talk a little bit about this and they might as well, right? As we continue to grow as a company, I think that the. We look into those things that are manual intensity, intense to, to, to and, and that not necessarily replacing human, but having them do more added value work. So I think that the fully autonomous invoice to payment piece is what really we are looking for. I mean from receipt to reconciliation. Right. From, I mean we, we talk about the fact that specifically apsc. I mean we have also a complicated chart of accounts as Damon has one too and multiple subsidiaries and locations here in the US and in Canada. So it is a little bit complicated which creates an additional challenge for AI. But looking at models and agents that can really take that agentic approach and do a process, an invoice with or without a PO end to end, right. All the way to the payment, collecting the approvals and some of that. And I think once we have that fully implemented EPSC and automated that will also build the use case for other things. I mean billing is one that comes to mind for us as well. I mean we, because of the type of services we provide, the different lines of business that we have, some are very heavily dependent on how we collect hours out in the field. Some are dependent on materials and services. The billing also has a structure that's a little bit complicated, but it's also another area that has an opportunity for AI to automate a lot of the things they've done by builders. I mean we have builders everywhere at PSC because of the complexity that we have in that specific area. And again, not to necessarily remove those positions but move them into different roles or and also not increase the headcount. Right. As we, as we add facilities across the U.S. i mean we are close to 200 now and, and we continue to grow. So it's something that I definitely look for and I think the AP will open the gates for other things but we really need to have that one solved where it's end to end and I think AI can provide a level of accuracy that's even higher than humans. In that specific use case, it's be more consistent. And also AI has the opportunity to, in that specific use case for AP to identify trends and duplicated invoices or charges that shouldn't be there and things that I think humans might not necessarily see. So there's definitely also, I think additional, not necessarily just replacing the human perspective, but adding value as well. I love that Elizabeth and Damon, anything different or anything that you would double down on what Alex said there about the AI use case that you think will deliver real ROI in the next two years? Yeah, I'm sort of in line with what Alex is speaking to like from a procurement perspective. We're really excited for the next phase with, with AI with Procurify like I want my team to be able to put a quote in and it extracts all of the data, creates the purchase request and then, you know, we'll train that model to then start to know, you know, this is going to this category or this is for this entity. Because even with the contracts module as well and the ability to like read the contracts and pull data out of it, I think that together anything that we can automate and again to save time and remove chance for error. Because even right now as they're manually inputting the data into a purchase order request, there's room for error every single day. So by, you know, being able to automate those, extract the data and then simply send it out for approvals and approval routes that are already built would save so much time. And then on the other end of it when it comes to AP and, and you know, receiving invoices, paying those invoices and then making sure that they have that three way match if they're able to, to do that automatically with hey, I've got this invoice now and this automatically matches with this PO and they don't need to search for it whether there's a PO or not, to Alex's point, because we have, you know, invoices that come that don't have purchase orders either, but anywhere that we can create efficiencies to, you know, decrease the manual intervention that's required today. Automate things so that people free up time to be able to use other, to be able to, you know, work on other strategic initiatives or something that may be, you know, sitting a little bit higher than, you know, manually inputting those costs. We're really looking forward to that. I love that. And Damon, anything to add on to that area, not a lot. I think they said it very, very beautifully. I think the only thing that I would add to that is if there is manual input, I would love AI to get to the point of being able to correct things and say, nope, this is incorrect, or this is the wrong, you know, routing or code or whatnot. And so I think, and I think we'll get to all of this eventually with AI here shortly. But having AI be smart enough to correct the human input and double check us would be amazing for those times where you can't automate 100% of everything. But yeah, I think I double down on both of their answers because I think it's spot on. Amazing. Well, that brings me to my final question of the panel. So we've had representatives here from education, from logistics, from trading, trading solutions, really looking for if you're advising another finance or IT leader, depending on your role today, tomorrow in your space, and they are thinking about what should they prepare for or how should they prepare their team for AI? Let's say they've been in a cave for the last while, they haven't taken a step, step forward. How would you advise them in your particular industry? What do you think they would need to know? Or it could be industry agnostic, but something that you wish you could have told yourself a few years ago. Don't be scared. Everyone's scared, right? And they should be. I get it, it's new, it's fear. Everybody fears anything. That's new. I was scared, I didn't want to use it. And so that would be my advice. Don't be scared to jump in because you can make a mistake and maybe it's not going to be perfect, but everyone, especially in technology, is going to be leveraging this. So don't be left behind. I'm a super early adopter, right? It drives some of my team crazy because I mean, my, my senior manager of cyber, he gets a little, I mean, I do everything on a secure way, but he gets a little bit excited about the things that I want to adopt, right. And want to test because at the end of the day I'm a geek, right? So it happens naturally. But I think one of the things that the major pitfall, and I've been doing this since I started working with technology and this is being a mantra for me even before AI was a thing, is that I try not to start with the technology. So it's really to look at a specific pain point. Just pick one measurable business challenge and ask the question, can AI Help me solve this, right? Instead of trying to pick an AI or a vendor that say they put AI, some AI flavor in their solution and then come in and try to make that fit your organization is really the other way around. That should be right. You have a business challenge, there's a gap. And the question is, can AI help me have a win here? And if it can, how? And then I think when you realize that when and you implement, that will help create momentum as well because you'll be able to see, okay, yes, I had a business challenge and I solve it. And that will help drive the cultural adoption because you can make a use case of that to your leaders and others in the organization and say, look, we had this specific problem, here's how AI solve it. But it's really starting with the problem. And if you try the other way around, I think there's, you can fail. I mean, in many of these tryouts, right? Just trying to put something in. And I think AI has that too, right? So first AI, when you engage it, you are amazed and at one point in that interaction is going to really let you down, really bad. So you go from here to here when start hallucinating and providing or making up some stuff. When you deal with ChatGPT and start making up some answers or things that don't exist, right? So again, when you start with the problem and you see how AI can solve that problem, I think that's very measurable and it really help drive adoption going forward. Alex and Damon, you have the last word of the day on this topic. So I agree with Alex and Elizabeth. I think start with the problem, the technology will follow. You'll find the right technology for the problem. So don't be scared, don't be afraid to fail. But as I tell my team quite frequently, buckle up because no one's going to die. People are going to think they are, but no one's going to die. And you can make mistakes and things can go wrong and you just have to pick up the pieces, figure out why something went wrong and keep going. But at the end, if you start with the problem and you identify your solutions and the processes that need to be in place, it's going to be a victory. It's just going to take some time and it's okay to have some failures along the way getting there. But yeah, buckle up. It's a wild ride with AI for sure. I love that. That's actually a theme I've heard, I think you say a few times today, is really having that kindness kindness and compassion as you're testing and iterating. Because to Elizabeth's point, there can be some fear, uncertainty or tentative feelings around dipping in or diving headfirst in. And that's really about going for it. Right? So to summarize what the three of you said on this answer, really about leaning in today, it's about leaning forward, working with your vendors, working with your team to really find those business challenges that could really benefit from AI. That's a really smart use of AI and a really great success that if you can actually prove that out, builds that momentum over time. And I know, Damon, in previous conversations you've called it really that trust fall, that trust fall for finance. And really being all of you today, being that frontier leader to lean in to drive that change. And we so appreciate you sharing your story with us today. And unfortunately, that's all we have time for our panel discussion. A huge thank you to Alex, Damon and Elizabeth for joining us today and for being so generous and insightful with sharing your story of AI readiness. So thank you very much. Thank you.