Helene: Hello everyone, and welcome to Procurement Game Changers. I’m Hélène Laffitte, your host and the CEO of Consulting Quest. On this show, we explore the strategies, innovations, and stories that are reshaping the way organizations buy and work with suppliers. My goal is to bring you insights you can actually use—straight from the leaders and practitioners making procurement a true driver of value.
Buying AI-based solutions is like navigating through fog—you know the destination could be groundbreaking efficiency and cost savings, but the path is filled with uncertainty. Is the algorithm really unbiased? Will the vendor still be around in three years? And what about compliance, data security, and scalability?
That’s exactly what we’ll explore today with our guest, Debo Lufadeju—a global procurement and project management professional with over a decade of experience across industries. Debo has led complex sourcing projects for SaaS, cybersecurity, and professional services, always with an eye for balancing innovation with practical risk management.
In this episode, we’ll tackle the big question: How do you navigate uncertainty when buying AI solutions—and still make smart, strategic decisions?
Debo, welcome to the show!
Debo: Thanks for having me, Helene.
Helene: Let’s get started right away. So you had such a rich career across SaaS and cybersecurity and professional services. What first pulled you into procurement and how has that journey shaped the way you look at New Tech like AI today?
Debo: So, I’ve been in procurement, for the past, 15 years. And I have an engineering background and an MBA as well. I think, being able to utilize both my engineering skills and my business school skills drew me to procurement. Procurement used to be called purchasing in the old days, and now you have, the term procurement.
Form a strategic sourcing, looking at things from a different perspective, having a holistic view, total cost of ownership, and really trying to, look at all aspects of the organization and, determine what value can bring in by procuring products, services, consulting, and so on and so forth. So, the idea of joining procurement or entering procurement allowed me to be able to use these skills, engage with several business partners, and be able to solve real life problems to be able to support the company’s bottom line and, growing future.
Helene: Yeah, of course. And people with engineering background and MBAs are very good people. I just can’t say that when it comes to buying AI, it doesn’t feel like the typical IT purchase. So, from your perspective, what makes AI a whole different ball game compared to traditional software?
Debo: With AI, AI is very dynamic. So, traditional software is considered static. It’s a deterministic. With AI, you have to think about the machine learning or the natural language processing, inputs. And because the outputs are very. Probabilistic. It’s not defined. You have to ask a lot of questions, understand what the business need is, understand whether AI can actually solve these business problems or not.
So, AI presents a lot of problems when we think about data privacy, security and if introducing these AI tools will support the roadmap for the organization. So, it’s a different problem for organizations to think about and determine if it’s something good to bring on board.
Okay, so you know, one of the big hesitations we hear is that AI comes with a lot of unknowns. Like the outcomes are not always predictable. That’s what you were saying. And the AI can be tough to pin down. So how do you think procurement leaders can make decision with confidence in that kind of a certainty?
So, I think, with the type of uncertainty that AI brings in, you need to focus on the business problem. What is the problem you’re trying to solve? And with this a lot of companies now are utilizing pilots, so small pilots of proof of concepts. Testing these tools out, very controlled environment to determine if they are gel or able to work with your existing environment.
Leverage. Open APIs determine if you can integrate these tools into your environment. And of course, with AI, because the tool is evolving and building capacity over time, you want to make sure that there’s a long-term approach to it. I don’t think organizations are bringing AI into use it for a year or two.
This is more of a long-term approach. So on the look at what your online long-term roadmap is and determine if, AI is a good fit. So flexible contracts are required for AI because you have to be able to pull out if required and also inserting performance metrics to be able to assess whether the model is biased, unbiased, it’s giving you the right results.
And also, just address the data concerns privacy concerns as you go along. So it is interesting to see that there’s so many things to think about before you even jump into AI, but I feel like still many companies are just jumping into it because it’s AI and it’s trendy, everybody’s talking about it.
Helene: And how can those companies make sure that they’re not indeed just acquiring it because it’s trendy, but because it’s a real business, it’s what are the steps that you need to take before you get even you even get there, right?
Debo: I think the most important question is why are you engaging in AI?
Why now? Is your traditional software or your current state, is it not providing the expectations you require. So asking all the why’s, the how’s, the when, the which, the what, again going back to measurable value, are you able to measure the value you’re deriving from this? New AI feature or tool you’re trying to introduce, IE reduce cycle time for procurement or generate MSAs sales or contracts.
Is it improving accuracy of your data in a certain way? Is it enabling your company to scale in future? What is the problem that AI is solving? And if you are able to tick all these boxes, then certainly bringing AI to a, into the environment is a no brainer.
Helene: Yeah, and it feels like the risks are bigger, like you have you were mentioning the bias, you, the data privacy, the ethics, even the vendor locking.
So how can a company put the right guardrails in place to manage those without slowing the process and the innovation that, that is so much needed.
Debo: Yeah. So, I think a risk-based framework is very important in a lot of industries. Now you hear, you, you hear a lot about third party risk management.
So, risk management has become a very key co component in procurement. Not only assessing your suppliers but assessing your suppliers. Suppliers, right? And, down that chain. So establishing a risk-based framework. For instance, bringing a tool that’s going to be on-prem and just maybe generating a document based off internal data may not be as risky as a tool that requires an LLM model that’s sitting in the cloud, right?
That needs to leave your internal environment to be able to utilize an algorithm sitting in the public crop. Cloud, right? So it, it all depends on the architecture, the it architecture of what you’re trying to do. If it’s sitting in-house, it’s less risky. If it’s going to be a cloud based or in the public cloud, of course there’ll be a lot more risk management towards that.
So there’s other things including a termination for convenience that can be inserted into contracts for you to allow to exit early. Also periodic audits, as we’ve said it’s dynamic. These models are changing, evolving. There is supervised data being inputted into these models. So, the model itself, the algorithm is changing.
There could be bias with the model, some ethical issues. So it’s important to have audits of the model periodically to ensure that you’re deriving the right value from it. Also, you can talk about data privacy, which is very important. There’s GDPR and other privacy policies that need to be adhered to.
Companies don’t want to be exposed and be fined for breaching these various statutory policies, so on and so forth. Other things you can think about is vendor lock-in. Most of suppliers or vendors are in it for the long game. They want to come into a new environment and they want to be there for five years, 10 years.
So having open APIs or solutions that can be integrated into other, either existing or future architecture is important so that you’re not tied to one vendor for the long term transition rights. Whether those, am I able to exit out do I retain my data? Who owns the data, so on and so forth.
These are questions that all need to be answered.
Helene: Yeah so what I hear is that it’s not only about the technical aspects. There are also some pure contractual. And also transition aspects that need to be taken into account because it’s uncertain. You need to be prepared for whatever happens next. And however that model will evolve in the future, because we don’t know yet, as you were mentioning, it’s not a static. So we see that, the AI market is pretty much buzzing with, a lot of players and new players every day. If I push a bit like start ups and niche players, but, those market that are exploding and booming, we know that those players, they may not be there in five years from now. And also the uncertainty is not only about the models themselves, it’s also about the ecosystem themselves. So how do you approach that vendor due diligence in order to make sure that you are not taking even more risks with the wrong horse, as we say.
Debo: So with AI, the due diligence is quite different from your traditional, looking at past performance, looking at financial data. Of course, all of those are still important. Determining what type of resources the company has, the capabilities, their pricing, so on and so forth. With AI, I think the technical team is the most important thing.
It is a very technically driven product. So what’s the technical team? What’s their experience? What’s, what industries have they worked in to be able to build this tool? You can scrutinize their security posture by running audits, like I said, data handling on. Doing some model validation and asking these suppliers to provide these logs so that you can run some tests to determine if the tool or the algorithm is actually providing the results it should provide.
And really looking at things like ecosystem who are the investors? Who are current customers? Can you engage their current customers and get some feedback or reviews of how the product has worked for them. But ultimately you want to have pilot projects, right? To be able to test the tools themselves and to see what results you can derive.
And if the results are positive, then of course you want to go ahead and engage the supplier possibly on a long-term deal. Yeah, because there’s a good chance that if it’s working and if the product has value, then they’ll be there in five years from now. Indeed. Yes. So that’s very interesting.
Helene: But a lot of those, questions are very AI related. Meanings that you need to have expertise in house. But the problem is that, most procurement folks are not AI expert, even though if they are, IT experts. AI is a completely new game. So how can those team, build the right knowledge or learn on cross-functional partners, like a data expert to really make those right calls?
Debo: I think it’s important for any organization that has resources to understand the core concepts. So, what is machine learning? What is natural language processing? What is generative AI? What is deep learning? What is supervised unsupervised data? So that type of stuff. But ultimately you want to have a leverage, a cross-functional team.
Procurement folks are not technical. Some are, but you’re not expected to be an AI expert, so it’s important to have that. Very strong internal alliance with the business team, with it, with the risk team, the data team model validation, and, importantly legal and any other teams that apply and really have that holistic view of AI and procurement really are the quarterbacks who are able to orchestrate, bringing the evaluation get their evaluation agreement signed off, negotiate the commercial risks, negotiate the contract and terms, putting all those pieces together. Also it’s important for organizations to be able to invest in talent, right? Having the direct. Procurement talent and providing courses, short courses for them on AI ethics data governance because technology changes very rapidly.
So it’s important for organizations to stay abreast and ensure that their resources are high fully equipped with all the new technology and the terms that allow them to provide the needed value for the organization.
Helene: I see a parallel with what we do in, with the consulting category.
It’s a bit different in the sense that you don’t really require to have a in-depth knowledge of each consulting. Type of project that you can have. But the procurement leader in that case of project remains more a facilitator of a process that has been agreed on, where you have several stakeholders that will participate in, legal and procurement of course.
And then the different stakeholders from the sponsor and the other stakeholders that are impacted in the project rather than really the knowledge owner or the one that needs to know what happens. But as you were mentioning at the beginning, I think that they’d still need to have that, deep enough knowledge of what is the category, what is it that they’re buying, and what are those terms that people are talking about and how a project is happening. And I think in that sense, it’s the same position. And then you rely on the people who know that can be internal, external, no matter, but they still need to rely on those people to bring the right insights on what’s right and what are weak signals that you need to look at and what are the red flags and the grid flags that you need to look at. So you mentioned that the beginning, those pilots and then proof of concept.So that seems to be the smart way to be in to step before your investor. In your experience, what makes a pilot successful and what are those like, weak signal warning signs. Signs that tells you maybe this is not a good, that’s not a good solution.
Debo: Yeah, I think for pilots, you want to have clear metrics like I said, cycle time cost avoidance.
What are the clear metrics or what problem are you trying to solve? Sometimes with AI tools because of the way they’re, produced there might be a lot of over customization. So you’re in a proof of concept and you find that the supplier is making so many tweaks or revisions to the existing software.
This customization might prove to be problematic in future. So those are things to look out for. And then again, I talked about open APIs and lack of integration existing systems. Very important. AI needs to be able to in integrate with your existing systems to be able to derive their correct outputs, human intervention, just being able to. Allow the AI do its work. If the humans have to keep interacting or getting involved then the process is not automated. It’s basically a manual process. So you need to try and remove human intervention as much as possible from it.
Some suppliers. As well are resistant to model validation. You ask them to provide logs of the existing models for whatever reason they’re not able to provide those data because it’s possible the model is biased or it’s hallucinating and doesn’t give you the right outputs. But it’s important to have an open conversation. It’s with your intended supplier. And if this is going to be long-term, then there should be no resistance to having, their model validated. But I think it’s important also to have specific use cases. So whether the use cases you’re trying to solve. So have a narrow scope and test these use cases in your pilot or your POCs, see what the results are.
And then most importantly, just having dedicated resources both on the customer side and on the supplier side. So not treating this as a hobby or a nice to have but really having a dedicated team looking to solve real world problems for the organization.
Helene: I think that’s a very good point, is ultimately what is important is to make sure that you spending enough time on that proof of concept.
And that you’re really capturing what kind of aligning what, where are the needs that you decided for the business at the beginning? And make sure that’s model answers, that and not all the nice to have that they can, you can see while you are exploring a solution. More and more procurement is being seen as a strategic partner rather than just, checking boxes and working on the transactional part of the work.
How do you see procurement stepping in into those roles on AI, for instance, but on all other very technical roles that come a lot with, some, those new business models and those new services that are emerging for organizations? Debo: Like I said earlier, procurement is a very customer facing role we have a panoramic view or a really right, wide view of what’s going on within the organization. We bridge the gap between the business, the risk, the legal, and really just put things together and put in very strategic, contracts or agreements that support the business and really just being able to adopt innovation early and responsibly.
So procurement should not be seen as the gatekeepers. We are just that conduit or that unit that supports the business in its entirety. So being just, like I said, able to putting all the right checks and balances bringing in commercials and the risk focus. So having a good tidy negotiated agreements, but also ensuring that all the risk components to the agreements are incorporated.
And also market intelligence. Procurement knows who the suppliers are, who the competition are. There might be similar tools that can provide a solution. It’s good to always engage procurement. We have an understanding of what’s going on in the industry and we talk to suppliers of vendors on a daily basis.
And so it’s just important for procurement to be able to assist. And quarterback all the initiatives going on within the organization and really build a good relationship with all the departments as well.
Helene: Yeah, no you are preaching the core. I’m all for having procurement, being part of those relationship.
Indeed. And I think also as you were mentioning that view of all the stakeholders, it’s also maybe a unique position where you talk to everyone and you in capacity to look at. What is contractor contradictory or what goes into the right direction? It’s not, it may not be the case for other functions, but maybe it would trust that client, internal, client facing step.
So yeah, I think that, that gives us a clear view of what AI is about, and how procurement can help their internal clients to choose the right model and how to lead that, how to lead that choice and that web process that can follow. But if you had, if you, if we want to leave one things to our listeners, one golden rules for navigating that uncertainty when buying AI solution, what would be that golden rule?
Debo: I would say a business value. So Forbes came out with an article sometime in August and said, 95% of pilot projects are failing. Really. So we want to understand what is the business value that AI is bringing and what problem are you trying to solve internally? Have a deep understanding of that challenge and where is the opportunity and what are the expected results?
So when you tie this investment to those measurable outcomes, so you’ve, you spent all this money on the tool and these are the specific outcomes or the problems you’re trying to solve, then, we’re ready to have that conversation. But a lot of what you see sometimes is that rush and that hype.
To have what you call a nice to have or something fancy. It’s not necessarily solving any business problem. So really take a step back and understand what the business impact will be, what the business value will be long term before engaging in an AI products. And of course, it’s important to remember that AI itself, there’s a lot of risk involved.
A lot of contractual requirements, statutory some for the organization as well. You have to protect yourselves. You need that legal support and have very strong clauses to be able to seal the deal. So really just have a very tidy business value or business proposition and you should be in, in good shape.
Helene: So to always keep your eyes on your business value and what are the needs at the beginning and how that solution can fulfill those needs. And that’s pretty much, that’s pretty much the golden rule. So thank, thanks Debo you’ve given us a roadmap that help us make sense of that AI puzzle.
I think that’s. We’re not the only ones that are looking at that phenomenon of AI. Wondering how we can also, in our organization, big or small, take advantage of that and indeed it starts with the alignment between the business needs and the solution. Having the right risk framework testing with pilot.
And seeing procurement not just as a gatekeeper, but really as a facilitator along the process and a strategic partner in innovation. So what’s clear is that there is a certainty that will probably be there, but there are ways to make sure that safeguard your organization, make sure that you have the right mitigation in place in order to really leverage AI in the best possible conditions, I put it that way.
Debo: Yes, that’s correct.
Helene: Perfect. Thank you.
That’s it for today’s episode of Procurement Game Changers. A big thank you to Debo for sharing his expertise and perspective on buying AI solutions. So if you enjoy this conversation, make sure to subscribe to Procurement Game Changers on Spotify, on YouTube, or wherever you get to podcast. And if you found today’s episode useful, share it with the colleagues who’s wrestling with the same challenge. We are here to spark the right conversations in procurement. So thank you for tuning in. Bye for now. And au revoir!
CLICK HERE to watch our previous episode
Consulting Quest 
