Video: AI and the Procurement Operating Model | Duration: 2644s | Summary: AI and the Procurement Operating Model

Transcript for "AI and the Procurement Operating Model": Alright. Hello, everyone. Thank you for joining us today for our webinar. Today's topic is AI transforming procurement. Is your data ready? And before I'm handing it over to our fantastic two speakers, just a little bit of housekeeping. So, yes, so for your best viewing experience, make sure you close any applications that might be running in the background. For your browser, we you we do recommend Chrome. If you want the best sound quality, make sure you use your headphones. If you have any questions, feel free to type them in the q and a box on your right side, and we are going to answer those questions at the end of the session. If you have any technical issues, feel free to reach out to me in the chat. And also last but not least, we will be sending out the recording as a follow-up, right after the webinar. So, yes, without further ado, handing it over to to you, James. Great. Thanks, Andrea. So, hi, everyone. My name's James, and I'm part of the product team here at Agiloft where I'm responsible for designing new AI features to offer our customers enhanced contract intelligence. Today, I have the pleasure of being able to moderate, this session today, and I'm delighted to be joined by our guest, Declan, who is an expert in the procurement field and has a wealth of experience in leading various transformation projects. So, Declan, welcome. It's great to have you join me today. Perhaps we could start and just give a little bit of background on yourself for the audience. Yep. Nice to meet you, James, and, thank you, actually love to invite for inviting me to this to this event today. So a bit of background about myself. I've been in procurement for about 25 years now, which is a long time. A lot of that was has been spent in transformation and digitization. I used to work for a company called ChainIQ, where I led the transformation and digitization team. I was part of the group extended group executive board, and, basically built, designed, and deployed our digital solution, that was I think, it won when I was there, it won an award from Everest, the research company for 2 years on the chart for the, it was called Star Performer if you like, and a lot of that was based on the digital solution that we built out when compared to our competitors such as IBM, Genpact, etcetera. So, after leaving there, I'm and now a partner at TPP Procurement, and we provide advisory services on how to deliver procurement transformation and introduce digitization and AI to to procurement. So thank you. Great. Well, it's great to have you join us today, Declan. I'm looking forward to our conversation today. So the topics that we're gonna cover today in this webinar, we're first going to discuss some of the key findings from a recent research paper and survey that's been carried out by TPP. Then we're gonna talk about why understanding your data is so important in the context of a CLM transformation project. We'll, of course, then talk about AI, and then we'll end with some practical tips for our audience to take away and think about. So hope that sounds, engaging for everyone online today. Declan, first question for you. Could you give us an overview of the research and the survey that TPP has carried out and perhaps start with, you know, who did you survey and why? Yeah. Basically, what happened is about 18 months ago, we were working for a client. They bought us into, redesign their operating model, digitize everything, deliver, like, regulatory compliance, and and that entire project involved, you know, some systems changes, process changes, and and in and recruiting a number of people from, like, junior analyst level right up to the actual, head of procurement. And when we were interviewing people, it must have been about 25, 30 different people we interviewed across the range of different experience in procurement. We found that, when we were asking people about how much you're using AI in your day to day job, it was actually pretty low. There weren't many people we were interviewing that said, yeah. We use AI day to day in procurement or we've always got it's embedded in our processes, whatever it may be. So for us, that was somewhat different from what we were hearing in in the in the news, in the media, etcetera. You hear this everyone's constantly bombarded now with AI. Does this, AI does that, but we were finding that the reality was that people weren't using it as much as we thought. So that's why we did the research. We sent it out to over a 1000 companies, and we got some feedback. And the second aspect of it when we were putting the research together was looking at the impact on the procurement operating model as well because we were making significant changes around the systems and the processes, the type of people, the skills that were required. And so the research also looks at some of the challenges, the company's gonna have to face, getting access to investment and and that kind of stuff. K? So when we did the research, we classified the respondents into 4 different categories. Yeah? Basically, the lowest level is non digital. There There were about 13% of the companies in there. Nondigital from our perspective is somebody who's still using mostly Excel, Word, Outlook, etcetera, SharePoint, a shared drive to manage their procurement function. Developers were companies who were or are, you know, they may have a p to p system in place. They may have a CLM solution or any sourcing tool, but it's not integrated end to end. Yeah? And they don't really have the right data management or data governance in place to start enabling AI and and really using it productively. Then you had advanced. These are companies who are on the first steps of using AI. They, you know, they may have it already in an application that they're using or they may start have started building out a chat a chatbot that they're using some large language model for the large language model LLM functionality for. And they're just beginning on that road map, but they've already started to lay the foundations around data management and data governance to enable that. And then you have market leaders. So these are companies that are actively using AI on a day to day basis. And we were quite happy with the the the figures that we got back because it tied in with a wider report that we benchmarked ourselves against with Gartner, and that was in the same the same same sort of area. So it was quite interesting to see our perception of how much people were actually using AI and was was correct, and it's not what you see with the hype in the media. There's a long way to go before AI is embedded into procurement on a daily basis. So one of the challenges that we also, identified was how do you put data at the core of everything you do? Because for AI, the only way AI works is if you use data and you you manage it correctly, you have the right governance in place, you create the right data, and then figure out how you're gonna apply that data to do 1 or 2 things. You can either with AR, you can either, deliver improved analytics or improved productivity. It's really that simple. There's not much more it does than that. You can do one of those 2 things for you. So that's a bit of an overview for the research we did. Yeah. Thanks, Declan. And we'll come back to, the the usage within procurement, a bit later on as as we move on. But, one of the things that you just mentioned, it'd be interesting to hear you talk about, you know, how we identify the need for data being central to, like, a transformation project. Right? Because we believe here at Agilof that taking a data first approach to contract life cycle management is essential to solving fundamental problems that an organization faces with their contracting processes. And what we're seeing in the market is that organizations need more than just a rich CLM feature set. What they're really looking for is for vendors who have a deep understanding of the data involved in contracting. And that data well, that contains a data taxonomy, if you like, and a model that covers the 4 main types of contract related data that we see here on this slide. So you've got things like contract content data, so that's things like key terms, maybe, causes that could be extracted out of a contract. Then you've got, contract process data, which could relate to things like contracts negotiation and cycle times. Then you've got that performance tracking data, which is very much focused on sort of tracking sort of key metrics that coming out of obligations that you might have agreed to in your contracts. And then finally, that contract decision support data, and that's really sort of data that an organization can use to drive, that business decisions forward and maybe start to automate some of their use to automate some of their workflows. And having a really good sort of understanding of that information architecture that's required to support your future business goals and objectives is going to be fundamental to achieving success here. How does this resonate with what you're seeing in the market today, Declan? Yeah. From the research that we did, yeah, that these these are definitely key topics that that that need to be enhanced. And I think like Keith keeps saying that all the way through it, and you can't underline it enough with AI is is how important data is. It's how AI learns what to do. It needs millions of documents or or or pictures or images or whatever it is to to to start to develop and learn. And, you know, in if looking at it from a procurement perspective, one of the things that I always find having worked in procurement for a long time and advising people on, you know, how to improve things now is the procurement cycle time. Yeah. It's a never ending bugbear for procurement people on, how long it can take to collaborate with legal, how long it can take to read through a contract, identify the right clauses, answer questions and queries around that, and anything that you can do, that focuses in on improving that cycle time and reducing and this is where, you know, AI with the productivity aspect of it can really take away a lot of those mundane tasks. Yeah? Once you start taking away those mundane tasks and you can create the I think one of the key aspects is what people maybe don't realize as well is when you start using some of these tools, you turn everything into data, and you have this data lake sitting underneath your underneath your tool that you can just access all this information instantly. And that gives you a real advantage in increasing that cycle time. That means you're doing you can do more RFPs, you can do more RFQs, etcetera, with the same teams. You can start driving competition. You're not spending, you know, 5 or 6 weeks negotiating the contract. You can condense that into 4 or 5 days, and that will really allow you to handle a higher volume of interesting work. Yeah? I mean, I'm not one of these advocates who thinks that AI is gonna take away people's jobs. AI is gonna make people's jobs more interesting because the mundane stuff is just gonna disappear. Yeah. 100%. It's about using the AI to work smarter, right, rather than thinking it's just gonna replace everything that we do. Okay. So going back to the research paper. So the the paper talks about putting data at the core of a procurement function, if you like. Perhaps you could tell us a little bit more about what a data centric procurement function looks like. And, you know, if you've got some examples that you can share, that would be great. Yeah. So I've got a really good example I can give you about how, from a data perspective, we completely transform the company's approach to, actually a CLM tool. But before I go there, I mean, because it it mean, it did significantly change the complete return on investment and why they wanted to buy CLM tool. It it completely transformed the approach. But I'll take a step back first and explain to you what we mean by putting data at the core of everything you do. Yeah? So, again, you can't underline and you cannot underlie underestimate, sorry, the importance of data for AI. Those who don't know anything about it, I don't know what people know or don't know on the call. But, like, for example, if you've used chat g GPT or Copilot or something like that, what the tool is actually doing is if you type if you type a question in, when you see the response, one word appears at a time. Because what the tool is actually doing is millions of calculations on the background based on millions of documents it's read on what it thinks is going to be the next word. So it's it's it's it's a massive amount of data that's been churned constantly to make these decisions for you. So the way you start well, first of all, you have to start creating data. The way you start creating the data is in the digitization aspect of it. First of all, you have to make sure you have holistic alignment with all of your processes. Yeah. You can't have, separations where you have a break in the process where you step out of the digital workflow. You really need it end to end and you need it holistic. Because if you start creating data or the wrong data, you're never gonna create the data that you want to actually apply to make your business better. And different businesses have different requirements on what kind of data that they want to create, whether that's from an e sourcing tool, a CLM tool, a spend analytics tool, supplier performance tool, whatever it may be. You need to really consider what data you're gonna create, how you're gonna manage that data, how you're gonna govern it, and have a plan for how you're going to apply that later on. Yeah? So once you've got your it's the first step of your actually creating data, what you realize is that starts to impact your operating model. We've already talked about the systems and the processes, but one of the key things for me that's gonna change in the future for procurement is you're gonna have data scientists and what they'll be essential. They'll be across all of your businesses, different functions, etcetera. Now whether the business itself actually ends up in 5 years time, having a data scientist team, like you have HR team or procurement team, etcetera, that are specialists, that you call upon, and they can come in and advise you on how to create the data, manage the data, and governance so you can use AI or, whether they'll be embedded within the individual teams, I don't know. But the use of those data scientists will become essential. The ChainIQ, we actually started to introduce because, we we were quite a bit ahead of the curve as I said with the some of the the the feedback that we got when we were developing this. And we actually you know, we had data scientists in the embedded into our my procurement team in, like, 2018, 2019, 2020 when we were trying to use different aspects of AI such as machine learning to speed up the process of any aspects of the procurement process. So once you've got the digitization and you've got the operating model, how do you apply that data? Yeah. So the example I'm gonna give you is for, a company that I worked with where they had it was a it's a large company. They had spend of over $10,000,000,000 spent around the globe in different regions, different divisions, diff a very complex organization. Yeah? And they wanted a CLM tool that went across the entire business. However, they took the initially, they took the traditional approach of saying, let's look at CLM. You can get productivity gains, from working with legal. You can increase collaboration. You probably improve the terms and conditions that you're gonna get because you've got templates that you're gonna be drawn down from the system all the time. You can have a nice contract repository where you can find your contracts. You know, you're gonna get these productivity across the whole of the procurement team and some of the legal team. You're gonna get some improved productivity and some improved risk management for the business because you've got better terms and conditions in place than you would have without using a CLM tool. So we we didn't go down that path. We took a completely different approach. And we said, look, you're gonna create a lot of data here. And you're gonna take tens of thousands of contracts, drag them through an OCR functionality of a tool, and create metadata looking at the pricing schedules, the the the the SLAs, the KPIs, the total contract value. You're gonna have good mapping of all the total, contractual commitment you have with your vendors over the next 3 to 5 years. You're gonna have all your start dates, all your end dates. This data is actually gonna be accurate because you can be confident that the Centimeters tools in the marketplace can deliver this. Yeah? So we said, right, what what can you do with all of this data? So we sat down with the category managers in a workshop and said, look, if you've got all this data for your different category strategies, what could happen to your category strategies? They were like, okay. This is brilliant. So we can now benchmark accurately, what we're spending for laptops across different business. I'm just using that as an example. There were many different aspects to it. Yeah? But nice simple example, laptops, we can identify what we're paying across different business units, different divisions, SaaS agreements. We can look at what maintenance schedules we've got in place that we may not need. We can look at coterminating large contracts across numbers of different business units, business different business divisions. So we can with this new analytical capability, we can completely revamp our category strategies. And we went back to the business and said, look. If the category managers only drive an additional 2% in savings, over the next 3 years, you're gonna save $200,000,000. And the category managers were willing to stand behind that that if they had access to this data, they would be able to deliver that. And if you work in procurement, you know, if you've got some good strategies in place, saving an additional 2% is is quite achievable. So all of a sudden, the business case that went up to the board transformed massively. It went from night, we can make procurement 10% more efficient and drive some, improvements in legal and manage risk a bit better. So we're gonna deliver 200,000,000 in savings, and we're gonna impact the company's p and l quite significantly. And all of a sudden, spending, you know, I think it was it was it was a 7 figure no. 6 figure. Oh my god. I can't do my maths anymore. It was in the it was over $1,000,000 a year contract. The the business case for it became a no brainer because they looked at things from a data perspective and how they could use what their data they created from AI and how they could apply that and completely revamp the value proposition that procurement could deliver. So I think that's a really good example of when you put data at the core of everything that you do, you start to look at things somewhat differently. Yeah. Great great example. Using it to calculate. I I guess the showing the the ROI there on that investment came much easier and and clearer. So Yeah. Really good example. Okay. Let's move on. So we've spoken about how important data is, to an organization who's perhaps thinking about transforming their procurement department and having a, you know, a good understanding of the information architecture required to support your future business goals is going to be an essential first step. If we think about the way that businesses are run today, everything really comes down to contracts. You know, without them, you simply can't buy or sell goods and services. You can't hire new employees, for example. And the challenge that we that we have when we look at all of our contracts today is that we're not really capturing the value that exists within each of them. And there's a couple of reasons for this. 1 being the sheer volume of contracts that might exist within your organization. And 2, most of that information lies in an unstructured format. So if you think, for example, you think about your legacy contracts, for example. And we we sit on this we sit on this mountain of information and this mountain of value that we're just not able to tap into. And this is where the AI starts to revolutionize the way that we can really access that information today. You know, it provides us with new ways to tap into this data and to help us find those crucial pieces of information that we're really looking for. So, for example, we're now able to ask questions about data. We're able to, like, summarize contracts. We're able to get those rich analytics to understand what's happening inside these contracts. Could I could could I just make a comment there, James? Yeah. Sure. I think when you look when people say talk about looking at AI in the industry today, one of the key things a lot of people seem to think AI has only appeared in the last 18 months. Yeah? And AI has actually been around for years. Yeah? Like I said, in ChainIQ, we were starting to use it 5 or 6 years ago. And it's different it's been around in different formats and different structures. And and and at the moment, what we're moving now is from, like, machine learning where you can teach, a computer or the AI engine to basically do repetitive tasks really, really quickly, yeah, that you can't do or make calculations that you can't do or recognize a picture or recognize a warranty clause or recognize an invoice. So it's been trained to identify the price on an invoice or something like that. AI was doing was delivering productivity gains. But since ChatGPT came out, that's changed everybody's perception of what they AI is actually capable of. Because, again, not everybody might realize is the difference between machine learning and generative AI is machine learning cannot actually create anything new. Yeah? So a machine could look at a a warranty clause and say, okay. I recognize that that's a warranty clause. Yeah? I recognize that that's the start date of the contract, the end date of the contract, and turn all this into data for you. But generative AI, you can actually say to the say to the AI, write me a confidentiality clause. Yeah? I've used it to write job specs. You know? Write me a job spec for a procurement manager, And this and it will automatically create something for you. So the difference in the industry today, I think, that's really driving the the the big investment, etcetera, and the big productivity gains and analytical capability, because, again, you can use the large language models to do some analysis for you, is all of a sudden people can see there's a huge difference between, like, machine learning, where it's just a repetitive task and actually creating something. And I think people really need to understand the difference between those two things and understand the impact that that can have on your day to day job. Yeah. Good point. Just to to reiterate on that. So I think, you know, AI has been around for a very long time. And, traditionally, we would have seen it, especially in tools like CLM tools or any sort of legal other sort of legal tech tools, for example. We were using it traditionally, to, you know, extract data out of contracts, out of clauses. And it was really focused on that data extraction layer. And now what we're seeing is those new models coming in, which are able to generate content, like you say, but we're now able to you know, execute on tasks. So the AI has always been able to execute on, like, simple tasks, if you think. But now we're able to really execute on more execute on more complex tasks and actually add much larger scale than what we've we've seen before. So, yeah, the technology is moving on quickly. But, Jacqueline, just with this all in mind, could you maybe provide what, you know, what are the key pain points that you're seeing for procurement functions today, and how might the AI be used to solve for some of these? Yeah. So I think we we already mentioned, like, cycle time already in the conversation that AI can significantly improve procurement cycle time and, you know, provide the ability for people to, you know, have more interesting work and have more interesting jobs if you ask me. Yeah. But another thing is that I've always found painful, in procurement. And this is, you know, when we're doing the research and people were, look, asking us to, you know, where could we implement AI? Where what what does this technology do for us when we're interviewing people? Is and if you worked in procurement a long time, you'll you'll be or even not maybe not even worked in procurement a long time, you'll have come across the problem of you're constantly querying data in a contract database or register reg or data contract register. Yeah. And invariably, everything's sitting in a PDF format. It's not necessarily structured correctly. You have to when somebody asks you for, can you tell me what's happening with the warranty clause, in a contract with HP, HP, from 2 years ago, you might find you've got 20 different contracts on the system all relating to HP. Then you have to find the right contract, then you have to find the right clause. I I mean, I've spent hours and hours just looking through PDF contracts and then finding out that somebody never uploaded the contract properly or they named it incorrectly or it's just not on the system. So that's quite painful. And from a contract perspective, AI with by turning all of that into data, really get can can really get rid of that problem for you. And, again, it's about creating those opportunities to work on more interesting things than you know, I can't remember how many hours I spent plotting through, contracts just trying to find somebody's query on what was the KPI or what was the SLA for this for this month, for this type of service. Yeah. Yeah. I'm sure a pain that many of our audience have have also felt. Right. Let's go back to a topic that we, you mentioned, you know, on the first slide of looking at sort of usage. You know, the in your paper research paper, we're looking at sort of understanding the usage of AI in the procurement functions today. So here at Agiloft, we have our own research team, and we also recently conducted a global study of procurement professionals. And this is this is, you know, what we're seeing on the side is some of the key findings that we found, and I think this kind of aligns with your paper as well, Declan, or the paper put out by TPP. And what we found is that many of our respondents aren't actually taking advantage of or currently taking advantage of these AI capabilities. So 34%, we're seeing do not use AI today simply just because it's not available to them in the tool. Whereas 57% have said that even if it is available, they're still not not gonna use it. Right? So, you know, Dan, are you are you surprised by this? Any explanation for why we think, usage in procurement teams seems a little low? No. No. It doesn't support. Again, based on the research and all those interviews I did last year, you know, that it doesn't really, surprise me. And I think it's probably driven by 2 different things. 1 is, a knowledge gap. Yeah? Look, I I've spoken to lots of people who use tools, but don't even realize that they've got AI embedded in the tool. Yeah? You know, Gartner call them intelligent applications. There's a number of them out there. There's there's quite a few out there. AI is embedded in the tool. You can use it whether it's just to, you know, with invoices uploading invoices into a system quicker than than manual intervention, something simple like that. Or whether it's more complex using chat GPT functionality or large language model functionality. There's I think a lot of people may have access to it, but they just don't even realize that it's there because nobody's trained them on how to use it or or or the the the business just hasn't really been pushing the use of it. The second thing that we noticed and we in the we tried to address in the research paper as well was around the investment. Procurement has to compete with other departments to get investment in AI. Yeah? Everybody wants to have, you know, a new tool or a new system implemented to get productivity gains. You have to compete up against high HR, finance, sales, marketing, etcetera. And unless procurement are putting together a good business case to get that to get them self prioritized, which comes back to my comment about putting data at the court can change that considerably, then I don't think I think that's part of the issue as well. Lack of investment and the knowledge gap means that it's just not being used as much as it could be potentially be used. Sure. And I I sort of suspect that if we fast forward 12 months from now, for example, I would expect, usage figures or to be, you know, much higher than what they are today just because I believe that, you know, the awareness is growing around amongst not just procurement professionals, but, you know, contracting professionals, legal, sales. You know, there's just this inherent expectation, I believe, that the tools that you're purchasing today for your business needs will have some AI capability built in. And that's you know, I think that's been driven that hype or, if you like, has been driven very much by this explosion of, GPT and just being able to just making it accessible to everyone. Right? Like today, that's the game changer is that these GPT models are now accessible to everyone to play with on your home laptop or even on your phone. So You're right then. The expectation has changed considerably. And there's an interesting fact that ChatGPT got to a 100,000,000 users within a couple of months. Yeah? It took Facebook, Twitter, Instagram, etcetera, years to get to a 100,000,000 users. Yeah? Chat GPT did it in a few months. It's one of the biggest it's it's the biggest impact from a a new product being rolled out, to get to a 100,000,000 users ever. Yeah. Worth quickest. So, interestingly, what we what this our, research also shows is that, you know, when AI is being used, procurement, you know, one of the top use cases or a couple of the top use cases, one being, data extraction or third party paper. That's no surprises there. That's been around for quite a long time. The other use case at the top quite high at the top there is, the u this use of natural language for searching and and finding contracts. Right? And so I thought it'd be useful, for our audience to perhaps we could take a quick look at how we approach this here at Agiloft. And so what we're seeing here is our user, in this case, Andy. You know, Andy's searching for a specific contract agreement within his CLM, his repository. And in this case, that agreement happens to be with a manufacturing company called Robust. He he goes ahead. He finds the agreement that he's looking for, and now he wants to ask a question of, the AI. And in this particular scenario, you know, we have a piece of faulty equipment, that we've purchased from robust manufacturing, and Andy wants to understand what his options are. So, we can see here, you know, he typed a question. The AI is able to identify and tell Andy that the seller is actually still obliged to repair this piece of equipment under the warranty at no cost to the buyer. And then Andy can ask, you know, a bunch of follow-up questions to find the relevant bits of information that he's looking for. So, you know, he's found the relevant clause text that he's looking for, and now he wants to know perhaps what the effective date of this agreement is. And then he can even go ahead and ask, you know, and confirm really with the with the AI, is this still within the warranty period? Yeah. For me for me, James, what's quite interesting about it is if you get this right when you when you deploy it, it democratizes a lot of the aspects of, getting access to to data across the business because, you know, procurement people don't need to sit there and show through contracts anymore. If you that somebody in the marketing department, finance department, if they want to understand what the warranty clauses in the contract that they've got in place or what the SLAs are or the KPIs, they can just if they've got a user license, they can just drop in here, and they can ask the question themselves. You don't need procurement's, input so much. It can deliver quite significant productivity gains. Yeah. And so, you know, what we're seeing here is a really effective use of how AI can be embedded into applications to help users find the information they're looking for by using natural language search and interact, with the system in in new ways. And this and you think you this could be applied in many different scenarios. Okay. Let's move on to now our last section. And this is where we want to provide our audience with some practical advice and guidance on what to do next if you're considering especially if you're considering taking advantage of some of these AI capabilities that are available in the market today. So, Declan, what are your thoughts on how to best enable procurement teams to utilize some of these capabilities, and what what are the key considerations that they need to be thinking about? Yeah. So lots of people who are, you know, trying to figure out how to use this at the moment. Yeah. I mean, basically, the in the short term, you can just start using intelligent applications. Yeah. If you're under pressure in your company to start using AI, you can either look at your existing tools that you're using and say, okay. Where is the AI in here that that's been embedded in it? Or, I'm gonna go out to the marketplace. I need a new CLM tool. I need a new resourcing tool. Tool. I need a spend analytics tool, whatever it may be. And start thinking about when you're selecting that tool, what AI is in there, and what productivity gains can I get out of it, and what data can I extract from there, and all these these kinds of access, all these kinds of, queries in your, evaluation process that you go through for selecting? And there's no reason, if you're quick about this and you have the budget, you can start using AI product productively on a day to day basis, in your department by the end of the year. Yeah. The spend analytics tools out there that you could buy that can advise you on, you know, savings opportunities, just by going through your spend data for you, and it's often what it's learned from other companies, etcetera. So, there are opportunities out there to start using this almost immediately, yeah, or as soon as possible. If you put a business case together this year, 2025, you can have AI embedded in in a certain aspects of your of your processes already, yeah, if you're not using it. Then longer term, that goes back to the research that we did. You know, you should have to start thinking about how am I gonna get everything working end to end digitally? How am I gonna align all the processes holistically with all of those systems? How am I gonna upskill the staff to make sure that they're using it correctly? How am I gonna inform the end users? Do I need to recruit data scientists? Are they already existing within the within the company, and how do I leverage that? And how do I put some plans together for the next 3 to 5 years so that 5 years down the road from now, AI is just completely embedded into the processes that I use. And I think procurement has a great opportunity here in comparison to some of the other departments to really deliver a significant change in the value proposition that they can give to, to their companies. Great. Good stuff. So I think we're coming to a close on the webinar today. Declan, I've really, enjoyed this conversation today. Thank you for for all of the insights. Andrea, I see you've come back on stage. I think I see a few questions in the chat. So have we still got time for some q and a? Yes. We do. And, yes, a few questions came through. The first one would be a lot of data is used to train AI. How do you ensure that the data we use that could be our company's IP is not mixed with other company data? Great. Who wants to take that question? Well, maybe I I can start and and just talk a little bit about how we approach this at Agilof because, you know, this is a real genuine concern that we hear from a lot of customers today. So, essentially, you know, we host a copy of, the LLMs, in this case, GPT 3.5, within our secure cloud. And that means that we're not, you know, we're not accessing this data running away. We're not we're not learning from this data, and this data is not shared, with others. So, you know, it's not shared with other customers, and it's certainly not shared with, OpenAI. And all of this data is really securely contained within our Microsoft Azure OpenAI service. Declan, anything else to add? No. No. It's just, again, this will come across it all the time, but there are, you know, the there are ways now definitely of making sure that you do not share your data. You know, I and I understand that that concern a 100%. You just don't wanna I I think it was a there were a couple of companies initially when originally, when chat GPT came out, we're dumping code onto chat GPT and and giving away, all their information, asking chat GPT to rewrite it for them and things like that. So yeah. No. It's definitely a legitimate concern, but, all the big organizations are addressing this now. So it should should sort itself out, but a good question to ask. Thank you so much, Bob. I have a second question here as well. That would be, I speak to so many software vendors today who all talk about the use of AI in their products. What else should I be looking for? Okay. Yep. Another good question. I'm glad someone's asked that one actually because, you know, obviously, AI, obviously, the topic today has been about pretty much the the uses of AI and, some of the benefits that you can to get out of it. But, actually, it's not it's certainly not the only thing that that you should be looking for, probably not even the most important thing that you want to be looking for. You know, first, I think you want to be ensuring that you're buying a platform that can scale for all of your needs. And so, you you know, recognizing that the needs that I have today, it could be very different from the needs that I might have, say, in, like, 12 months' time, and thinking about, you know, your data needs as well, for example. And and that's when you want to make sure that the products that you're looking at have the flexibility to be able to handle, you know, all of your data needs. You know? Because, for example, if you're looking to identify or add specific metadata fields to the data model that might be important to you, not only just adding new fields, but also controlling who has access to this data. Right? And then you also wanna think about what what do you you might extract that data, but then actually you might wanna think about, you know, who who else wants to use this data or where else do I need to access this data? And this is where thinking about sort of the integrations that your this system offers into other, perhaps, downstream applications like procurement systems, you know, thinking about those downstream applications and and the people that wanna access those. So for some people, some of your users might be quite comfortable using the CLM or might be having a need to access the CLM every day, but there are gonna be other users in sales and procurement who wanna access that information in other, other applications. And so ensuring that you have, you know, the solution that you're looking at has the right integrations and making not just the right integrations, but makes those integrations really simple for you to take data out of one place and your contract lifecycle manager, for example, take data out of a contract record and send that data to somewhere else where that information might want to be consumed is going to be really important also. Yeah. I mean, I agree with that as well. It goes back to my whole thing at the start. You have to consider what you're gonna do with the data. That's not just the the AI itself, but what data you're gonna create. That's something that people are not evaluating at the moment. I don't think not widely anyway, when they choose a new tool or a system. K. Thank you both. And one last question before we close, where can I access the research and survey that has been talked about? So, we will be sending out a follow-up email in the next few days with the links to both surveys, and you can also find them, on the website as well. So, yeah, with that, thank you so much, Declan. Thank you so much, James. I hope this was an insightful webinar for everyone. Thank you for joining us, and, see you soon, everyone. Thank you. Thank you very much, guys. Cheers. Bye. Thanks all. Bye bye.