Ep 8: Supplier Management AI Use-cases feat Jarrod McAdoo

Welcome to the Love Procurement Podcast, where we bring you a fresh and lighthearted take on all things procurement. Join us as we dive into the fascinating world of procurement with a twist of humor and a sprinkle of fun. Hello, and welcome to this episode of the Love Procurement Podcast. My name is Kelly Barner, and I'll be your host for today's conversation.

Today, I'm glad to welcome Jared McAdoo back to the show. Jared was one of the very first guests we had in season one. He's a director of product marketing at Ivalua. And hi, Jared.

It is great to have you back. It's great to be back, Kelly. I'm glad to be back because I thought maybe I did something wrong. It's been a minute since we've had a chance to talk.

Well, in that minute, you'll have to tell me, we usually ask folks at the start of these conversations, what is something that you love about procurement? In that minute, maybe you found something new to appreciate. What's something that you love about working in procurement? Yeah.

I think one of the things that I love about it, I could probably go on a whole podcast about that, but let's talk about the frame of AI now because everybody's talking about this. So what I love about procurement is sort of the recognition now that we're getting with AI that how do we use this to take stuff off of these people so they can think because they have such a knowledge, such a skillset when it comes to strategic relationships. you know, which honestly at certain times during my 30 years in procurement, you didn't exactly feel the love that you're starting to now.

And especially within a context of something like AI, you know, it sort of makes you proud. Absolutely. Now it's interesting that you bring up sort of the frame of AI, because that connects perfectly to what you and I are going to focus on today, which is supplier management. Yes, as an overall practice, but directly connected to some of the data quality challenges or opportunities, depending on how you want to look at it, and maybe how AI can help us with some of that.

I thought you had a fantastic analogy that you used when we sort of connected to prep for today, which was the idea that AI is the jet engine and supplier data is the jet fuel. What does that mean to you in terms of the current state of supplier data? And I don't know, maybe how fast our rockets are able to go. Yeah, I think right now, if you look at, you know, you need good jet fuel to run your jet engine.

A lot of organizations in the state of the data overall, we're running marine diesel. You know, it's that level of what we're trying to put through it. You know, just because that, you know, AI doesn't solve a lot of, doesn't solve every problem. Yes.

It's good at exposing those insights, but it's not going to necessarily fix the fact that you have Acme Supply Company, but then you also entered a duplicate that's Acme's supply company. And then you added a third one in there that says the Acme Supply Company. It's not the greatest thing about figuring out some of those nuances. And it doesn't necessarily create information.

Like it'll tell you if you're missing something, like, hey, you don't have a certificate of insurance. But if it's not out there to be scraped, it's not going to be able to create one. So, you know, understanding, you know, some of those items as well as how we can. It's AI sort of like an individual.

Try to put it in the best position to succeed. And I think you need to understand about what it is and what it isn't. It's not everything to everyone, but it is something in a lot of areas. So just understanding what that something is, I think is extremely important.

And if you do that and you fix some of those things to put it in the best position, then we're more away from the diesel towards that jet fuel. Yeah. Now, it's interesting because there is a lot of excitement around AI. We're putting a lot of eggs in the AI basket.

It's going to be able to help us fix all these problems. Do you find that that enthusiasm sometimes sees teams overestimating either what the AI can do, or maybe where the line should be between, okay, this is what AI is really great at, but this is where we need to bring in the people to do some of the more foundational supplier data work. Do you see those two things clashing? I do.

And I think it's just where we're at in the market, you know, as far as, so everybody's positioning the AI in every aspect of your life. You know, I have AI embedded in this, and everybody's trying to get your attention because you're hearing it all over the place. So we have a tendency to try to make the sexiest, most complex problem we're going to solve with AI. But the truth is you get a lot of the value, the immediate value, just as far as ROI, but the immediate value on your people and focusing on a lot more of those non-complex items that are easier for AI to solve, number one, and really have a benefit on people's lives that are really tangible, that you don't have to say, hey, I'm going to solve this complex problem you probably didn't know where you had versus, hey, you know, these annoying nine things that really keep you from getting to those important strategic items.

How about I take them off your plate? Yeah. And, you know, but again, it's like I mentioned to you is it's AIs everywhere and it's everything. And we're all trying to get the consumer's attention by solving the most complex problem.

And I think we need to take it a step back to say, here's how we can solve a hundred basic problems quickly. Yeah. Now, speaking of taking it back, if we want to start improving our supplier data, the first thing we have to do sounds deceptively simple, and that's decide what a supplier is. Now, obviously we have a lot of procurement and sourcing professionals listening into or watching the show, not insulting anybody.

We know everybody knows what a supplier is, but from a data standpoint, what are some of the different definitions or data points that might factor into that question of what is a supplier? Yeah. And I think procurement's good about knowing these answers, but they don't think about it. So let's just think about it this way.

like the normal procurement professional goes through this analysis and figures this out without even knowing they're doing it. So I'm a category manager. I'm looking at my spend. So I'm looking to understand if I spend this amount of money with this supplier, what are all those subsidiaries that belong to that supplier that I spending money with What my total spend picture And I want to at it as a company level because that where I wanna negotiate because that where I have all my leverage So they're already thinking at that highest level just because they're trying to get the most leverage for their spend.

And we try to think of that as supplier as well as I'm going to negotiate and contract at the highest level of the supplier because that way I can get my whole leverage. I also need to look at a certain compliance level at that highest level, like beneficial ownership, those type of things. So they're already thinking about those, but then you might have a direct material buyer. It's very, you know, it's very critical in the parts they're getting.

It's very critical to their internal processes, to keeping the assembly lines running. But they're thinking, where am I getting that part from? What is the site and what country and what risks they have there? So they're going through a lot of these thought processes as well.

Also, in their mind, when they're looking at supplier management or they're looking at contracting, they're defining these type of items. And now they can just bring that information back into the data structure, start setting them up that way. Then you get it from more of a tribal knowledge to the more ingrained items. And again, that gives you something you really start having a good basis to fuel an AI.

Now, as we think about how what we do today is going to factor into future AI-related work, we probably need to revisit the data that we're gathering at the point of supplier onboarding. So in the past, we probably took all the data we needed to make sure they got paid, to make sure that a contract was complete, to satisfy the fields in whatever system we're using. None of that was necessarily forward-looking, even if it met the needs of the moment. If we think about what we might want to do with supplier data in 12 or 18 months, even three years, given the rate of how AI is changing, how should we be rethinking what we gather from new suppliers?

Yeah, I think the sort of the challenge that I always see on gathering information from suppliers, or at least the challenge I challenge myself when I'm thinking about it is, it's easy to say I'm going to grab a bunch of information, even though if I don't know what I'm going to do with it yet. But there's a cost to that. You know, if you want to be the customer of choice, don't waste people's time. Just get the information you need.

But also take note of what you've already collected. Like, for example, if I collect a quality policy, because it's part of my onboarding process in order to provide this, I'm going to look at these, or if I get a CSR report, if later somebody asks me the question, do they have an SBTI commitment? Start to understand is, okay, normally I'd go out and ask the question. But now can I ask the question of what I already gathered?

It's probably in that CSR report. but I probably figured it's easier for me just ask the supplier than read a 50 page report. And again, those are the type of items too. When you look at it is, Hey, I collected a bunch of this data.

Why don't I go interrogate it? Great opportunity for AI because it's something it does well. Interrogates large amounts of information at scale to sort of pluck those things out to get those. So I think we can start gathering information that we only need better interrogate that information and then determine if there's a real gap where I need to ask again.

You know, because just because I can't ask and I can't get doesn't mean I should. Absolutely. Now that's gathering data. Let's talk a little bit about cleansing data.

You know, I think it's, it's funny. So before procurement, my plan was I was studying library science and I did an archival studies program where I had an internship one semester that taught me why I did not want to be an archivist. I was more focused on wanting to clean and neat than to preserve every single little scrap of paper. I think I came to the realization in that moment, okay, this program is not for me.

But I'm sure there's a lot of similar things that we would struggle with when we start to cleanse data. Because on the one hand, you can't necessarily keep anything and have a usable data set that makes any sense. On the other hand, there's all of these decisions of what is it safe to get rid of? Where are we really losing sort of tribal knowledge?

Where are we damaging our future ability to answer our own questions? How do you balance this process of cleansing and streamlining with genuinely wanting to preserve a history associated with that category or with that supplier? Yeah, that's an interesting challenge. I mean, in my years in procurement where we've had to address data hygiene issues, that was always one of the biggest challenges is, hey, I have 20 of the same supplier across all my business units.

Yeah. And we should get to one, but people were worried as I'm going to lose all my transactions there. Like that's, that's history. That's important.

And I know those were discussions when we had with our R and D team is, Hey, that information's not only important just for the procurement professional in the business, but it's going to be important later. Like for example, if I consolidate a supplier and I retain all the sourcing events that they participated on, the best way to discover a new supplier on a sourcing event is to look at who finished second five years ago. Because I look at it this way is we had a supplier, they participated, they were competitive, they were compliant with the process.

I know a lot about them, even though they didn't win. And now I can see if I want to invite them back. So I think there is, you know, sort of that challenge that you have on there is, number one, you need to look at solutions who can preserve data, because we didn't always have the option back in the day to do that is like, if you clean it up, you're going to lose it. but then also decide is what data is worth saving.

A sourcing event is probably worth saving. Payments are always worth saving. You know, what is the type of information that is not only going to inform procurement, but AP, you know, those other teams as well to get on that. So I think it's incredibly important to look for solutions to your problem that'll let you preserve those.

But then also to your point, you can't keep everything. because you know it's like anything in your house well if you haven't used it in seven years are you ever going to use it when you're getting rid of it so those are good questions to ask on those but you should definitely make sure that you have the opportunity to make that decision for yourself that a technology doesn make it for you yeah now I also know that you have experience on both sides of the sort of indirect direct divide of spend management Is the supplier data management challenge the same?

Is it similar with just maybe some tweaks depending on which side you're on? How common, to what extent can we just say, oh, I'm trying to do a better job managing my supplier data and it doesn't matter what suppliers you're talking about versus unique considerations that you would need for indirect versus direct spend? Yeah, that's a great question. And I'm going to try to answer this without, you know, sometimes I'm going to give you this answer and I don't mean it.

It makes it sound like indirect is less important than direct. And I don't mean that at all because they both have their unique challenges. They can both shut down your business. But like, you know, the way I look at it is a lot of the indirect material, you are ordering sort of a commodity.

So, you know, and I don't mean to be insulting with what I'm about to say as an analogy, but there's a certain aspect when you go to a McDonald's, whether you're going to a McDonald's here where you live or a hundred miles away, you get that level of consistency. And when you're buying those type of commodity type items, you can manage it at that highest level. And, you know, without having to go down to the individual city. You trust an organization, especially if you have a good indirect supplier, they're going to give you the right commodities and you can maybe manage that in more of that company level that we talked about.

Now you get in direct materials, it can be entirely different. It's just a different perspective where you might want to get down to that site level, that factory level, whether that factory is in Asia, whether it's in Eastern Europe, in the United States, you have three. So it's that different level of management. it's still amount of work you're going to go in.

You're going to work with those suppliers. But when you start looking at the data structure, how far down do I need to go? What level of contact do I have to need? Like I can manage a Granger because they have strong processes.

I don't need to get into their business at every counter. But if I have a critical manufacturing part where there's a critical process being done at this factory in Eastern Europe, I need to know that. And I need to understand that. Plus, when there's a risk, you know, if I see a risk crossing Eastern Europe, geopolitical, those type of items, I need to understand those type of items.

So it gets into that perspective that there is a difference just because of the understanding of what level I can get most comfortable with and what makes sense to do that, if that makes sense, what I just said there. Yeah. Now, we kind of started talking a little bit about AI, and I appreciate the fact that we've gone off into process considerations and relationship considerations and having intent around the data. But let's bring it back to AI to start sort of closing this loop.

You did make the comment earlier, which I think is incredibly important, that we shouldn't just, okay, we're going to bring in AI. Let's do the biggest, craziest, most complex thing that we think this AI can handle. Are there any examples of small or appropriately scaled AI supplier management use cases that procurement can not only just get their feet underneath them and start working in that way, but frankly can also deliver immediate value without some of the complexity that might otherwise hold them back?

Yeah, this is one, and if people who know me, they're going to probably say, oh, he's bringing this up again. But here's a real reality-based use case. Anybody who's been a buyer who's searched for suppliers and onboarded them, they probably at some point had to request a level of insurance, that a supplier had to have a minimum level of insurance. And they would certify they did, and then you would require them to submit a certificate of insurance.

And I guarantee most buyers are probably like me that I got them and I put them in a desk and I never even really looked at them. And we have some of those use cases when I talk with our customers, too. It's one of those. It's I hope it's never an issue because I really don't know.

Yeah, that's a great case. Just right. There's an example where AI is really good about you going through this process and it's absolutely necessary. Your insurance group will tell you it's necessary.

So legal. AI is real good to say, I got the document. I've stored a document. The metadata is there.

I've read it. It complies. This is something you just weren't doing before or you didn't have time to do every time that came in. But now it's a great use case for AI.

Just one of those, again, those nuisance type items that would either take your time or would give you an exposure because you didn't have time to do it. And other documents are like that as well. I mentioned, you know, you can get a CSR report. That could be a very heavy document.

And if you're looking for some specific information, and that's a great way for AI to interrogate it. You know, what are the SBTI commitments? What are the dates they're committing it on? Those types of things you want to get on it.

So I think, again, it's not sexy when you're trying to stand out and marketing it. But from the standpoint of those are very impactful things that you can get day one. Yes. On those.

And, you know, I think that's where you start building up. And And again, just look who I mentioned, you know, that impacts procurement and impacts legal that impacts your insurance group. Just on that one example who would say, oh, this is great. Now I have some comfort.

Like I can sleep better at night now knowing at least somebody's put some eyeballs on it, even if that somebody was an AI. Yeah. I think that's an interesting point because I would probably tend to think of supplier data management as sort of a procurement housekeeping task. It's something that needs to be done.

people are trusting us to make sure that a system they might access with supplier data in it, it's going to be complete, it's going to be reliable, but maybe it's bigger than that, especially as we start to think about bringing in AI and even making more information than we've had available in the past. To what extent should procurement be managing these projects as cross-functional sort of business projects versus something that procurement needs to do within their four walls to meet expectations.

How should we be thinking about that? Yeah, I think we should be thinking about it as a cross-functional team because, again, we just mentioned in that most simple example, the teams you get impacted. So if we think about a supplier who gets onboarded or we think about managing a supplier, I mean without really knowing it we have a cash flow that are tied to it We have payment terms We have all those type of items that impact other groups There this expectation from legal that if we build a compliance process we doing the necessary compliance checks whether it on a politically exposed persons list or whether we just looking at you know any of those lists that we have to check for those So there's our people involved.

And as we shift the AI, it's good to make sure that we have people involved so they understand this is our process. This is where we're employing AI. This is where we have a human in the loop. Because, again, AI is never going to be held responsible for making poor decisions.

You know, we also need to make sure that we're demonstrating that compliance. And it would be good so everybody knows how we're doing it, what the process is. And, you know, legal in the past might have said, well, I'm too busy to do all this. I'm going to build a process you need to verify.

Under this process, maybe they can be more involved. They can have more confidence in the system because it's not a big time drain from what they need to do as well. And you can really start bringing them into the function to not only be part of the process, but feel like they own the process as well. So doing all those type of things, I think, is a great opportunity to make this cross-functional because it eliminates some of the fear of workload we had in the past for all those teams.

And it's a very real fear, right? There's all these new things procurement's trying to learn. They're not saying, oh, we're going to lower savings or spend management targets to give you all a chance to learn a little bit more about AI, right? No one is saying that.

So we're all learning on the fly and we're figuring out all these new systems and we're trying to apply them genuinely to the best effect. Jared, we always like to leave people after these conversations with something practical that they can maybe take back and put into practice right away. If we were to think about something practical that sits at the cross-section of all the things we've talked about today, so certainly AI, but supplier data management, those supplier relationships, even the cross-functional relationships that are coming to bear as these systems become more and more useful, what would be something that you would recommend people think about, check into, take a review of in, say, the next quarter before they try something a little bit more complex?

with their supplier data, whether it involves AI or not. Yeah, I will say this is in my 30 years being in procurement and in business in general, you was always here to say that garbage in, garbage out. Yes. And you know, that has never changed in those 30 years and it still applies today.

So, you know, all this technology you have out there that even today with it and even with AI on top of it, you still have to understand that if you have garbage in, you're still going to get garbage out with that. So there's some basic things you need to do always. Like take stock of your supplier data. Understand how you've built your supplier data.

Understand if you are managing supplier hierarchies, do you need to manage supplier hierarchies? So again, it's an understanding of those type of things. And you could probably do it by look at your top 20 suppliers, top 20% of your suppliers, because they're probably driving 80% of your spend? Are those the ones that are important to you?

Or is it the other suppliers who are driving your transactions? What's important to you to driving business? Am I looking to drive down transactional work? Or am I looking to strategically manage spend?

And then look how you've built your supplier data. And that'll tell you if you have a problem. And it'll tell you how big the problem is. The last thing I want to do is, I think we all have an opportunity to improve our supplier data, but not all of it needs to be as improved as what we meant, as I mentioned in some of these examples, to run your business.

There's a right level for you. So determining what that is, I think, is extremely important because then as long as you know where you stand, you can manage other issues because there's always going to be other issues. It's when you have a blind spot that you think your data is good, that you don't know, or if you think it's bad and you don't know why it's bad, I think that's the challenge. So if anything, I would have people go back before they spend time investing in AI, do this as understand is AI is just going to amplify my business knowledge.

How do I feel about my business knowledge as far as my structured and unstructured data? And then I'll help you make some decisions going forward, you know, on the tools you select in the timeframe you select it. And to your point, the use cases you select on these, you know, if you go for the most exciting one or the one that's going to pay off the quickest. And I think a couple of the things that you pointed out are particularly important, partially because at least so far, AI can't do them for us.

One is prioritization. Only people right now understand the total big picture and what is the most important for what reason might not even be related to the biggest contract by spend. It could be something entirely else that we understand as humans that an AI is not going to understand. And then the other thing is that critical thought.

Is this data good enough for the thing that we're trying to do with it? I think that's much more appropriate for a human to think about, maybe debate, right? A little bit of multi-human back and forth on that, especially if we're looping in suppliers and people from other functions. Super interesting, Jared.

I appreciate your coming back. Clearly, it wasn't poor performance that led us to skip from season one to season three, because these are all really, to me, realistic, but totally contextualized things that people need to be thinking about with their supplier data, if they're using AI or not, but making sure that quality is at the right place. So thank you so much for coming back and sharing your perspective on this. Thank you, Kelly.

It's always a pleasure talking to you. Absolutely. And of course, thank you to everybody who has watched us or listened in to this episode of the Love Procurement Podcast. Explore through, listen to other episodes.

And of course, in a couple of weeks, we will be out with a new one. Thanks so much, everybody. Thank you for listening to this episode of the Love Procurement Podcast. Join us every other week for fresh takes on all things procurement.

Remember to subscribe, like, share, and rate the show so we can expand our community and share the love. Visit iValua.com slash podcast to access all episodes.