Season 2 Episode 3: Leading with AI in procurement

Evolving role of AI

As supply markets grow more complex, artificial intelligence (AI) is becoming essential for procurement teams to manage data, streamline processes, and minimize risk.

“The need has never been larger and greater to use AI to help manage all the complexity that's in the supply markets,” says Pierre Mitchell, distinguished research analyst at Spend Matters.

In a recent episode of the Smart Business Buying Podcast, Pierre broke down the research findings from Spend Matters' AI Leadership Guide for Procurement. He shared insights on how organizations are applying AI today, the challenges they face, and the skills needed to drive success as the landscape evolves.

As Pierre states, "It [AI] is a fundamental sea change that is happening."

Pierre focuses on three key areas for procurement leaders:

• Understanding AI's current and future applications
• Navigating the challenges of AI adoption and starting the journey
• Developing new skills for an AI-driven future 

Transcript:

This has been lightly edited for clarity. 

[00:00:07:09 - 00:00:42]
Kelly:
Hi there and welcome to the Smart Business Buying podcast. I'm your host, Kelly Rogers.

In today's episode, we'll be exploring the evolving role of AI in procurement and how teams can get started on their own AI journeys. Joining us today is Pierre Mitchell, the main analyst behind the recently launched Spend Matters AI leadership Guide for procurement.

We'll put a link to the guide in our show notes in this guide. Spend Matters drew on direct feedback from 350 chief procurement officers and their own research to look at how AI based tools are reshaping the industry. Pierre, it's great to have you here.

[00:00:42 - 00:00:44]
Pierre:
It's great to be here. Thanks for having me, Kelly.

[00:00:44 - 00:00:47]
Kelly:
So to begin, could you briefly share what speed matters is about?

[00:00:48 - 00:01:17]
Pierre:
Sure. Yeah. Spend Matters is a tech analyst firm. We do research and advisory, and we focus on procurement and related areas primarily, certainly in procurement, but in related areas around supply chain and finance and anything that kind of touches procurement. And we focus on helping companies understand the impact of digital on their operations and making smart choices given the, wealth of technology, overwhelming amount of technology, you know, that's out there to potentially help them.

[00:01:17 - 00:01:54]
Kelly:
Absolutely. Well, Amazon Business is incorporating, you know, machine learning and generative AI into the buying experience for a lot of our business customers. You know, for example, we found ways to make some of our search functionality more intuitive to better sort of anticipate the needs of a buyer. We're also working on refining some of our analytics capabilities to deliver more of these spend insights right back to the customer.

But we know that across the procurement industry, businesses of all shapes and sizes are thinking about incorporating AI in different ways. So I was hoping you could give us an overview of maybe how AI has been used in the procurement industry so far.

[00:01:54 - 00:02:18]
Pierre:
Sure. I mean, AI is an old topic. I was doing some AI development work. I mean, literally like 30 years ago. A lot of the technology has radically changed, since then. But if you look at what's kind of happening, you know, within procurement organizations, a lot of times they're using AI that they just don't even, see or realize is I sometimes it's kind of an older form of AI.

[00:02:18 - 00:03:29]
Pierre:
So for example, the area around auto classification and helping kind of understand your messy spend data so that you can gather it all in one place, analyze it, make sense out of it, be able to kind of, you know, develop, spend cubes, if you will, your spend by supplier, by cost center, you know, and by category so we can better kind of, you know, manage it.

So, that's an area where there's a lot of more traditional AI, often called discriminative or non generative AI, but a lot of that, you know, is quite often behind the scenes that you don't even see. But it's there, taking, you know, creating value. What's happened though recently obviously with AI is it's really taken AI to the next level, right in terms of the, adoption and consumerization around using things like large language model based chat bots and copilots and really, people just starting to get familiar with, some of the tools they're starting to now get used within a procurement context, formally and sometimes explicitly, but other times often still behind the scenes, to maybe augment and build upon some more traditional AI that I was talking about before.

[00:03:29 - 00:03:59]
Kelly:
Absolutely. And at Amazon Business, we're investing in a lot of those areas as well, really, driven by our customer obsession and what our customers are doing. I'm hoping you could give us a better sense for, sort of the different levels of sophistication of AI across the industry. You know, sometimes people perceive the procurement industry as maybe an industry that's maybe lagging on, hot trends and staying up to date.

So give us a sense for sort of the different levels of sophistication. Are people really jumping on, some of these new technologies and trying new things?

[00:03:59 - 00:04:09]
Pierre:
Well, to your point, I would actually agree with you in terms of procurement has historically been somewhat of a lagging function and, has not had a lot of investment, particularly around the digital space.

[00:04:09 - 00:04:30]
Pierre:
And so, you know, a lot of the real advancements are, typically happening on the sell side and in the consumer space. So in B2B and procurement, quite often it is a bit of, it's lagging. However, the need has never been larger to, you know, and greater to use AI to help manage all the complexity that's in the supply markets.

[00:04:30 - 00:05:35]
Pierre:
That in some senses is even more challenging problem than, you know, focusing on the customer. So it's kind of like that old quote about, you know, Ginger Rogers and you know, Fred Astaire. You know, it's like Ginger had to do it. Fred did, but backwards and in high heels.

And so kind of procurement is almost in that world of having to constantly support, what's happening with those who own the budgets in the business, but also with the external customers, as that gets passed back and they have to know what's happening in the supply market, know what's happening with costs and tariffs and all the market intelligence around that and supply bases and all these things.

It is very complex. So as we think about the level of sophistication, there's kind of two dimensions to it. One is a there's a technical piece. Right. Which is we've gone from basic machine learning, as I was describing before around Nam or that traditional, AI to more of the generative AI and using large language models to generate content and to, support some of those use cases, as many organizations are starting to get into it.

[00:05:35 - 00:05:58]
Pierre:
And mostly everybody is familiar with, you know, some of the chat bots around ChatGPT and just, you know, some of the different, chat bots that are out there. But as we then move into then the genetic world. Right. And how do we use, agents that use generative AI to kind of help, orchestrate various processes and do a lot of these tasks?

[00:05:58 - 00:07:36]
Pierre:
That's kind of what's happening right now in terms of that technical sophistication of taking that generative AI and now baking it in into more, robust, scalable, complicated, more B2B processes. And that's where, you know, the the current focus is right now. But certainly there's a lot of just basic blocking and tackling.

But on the business side, maybe the easiest way to just think about it is one, just our personal use of things like, you know, the chat bots that are out there and just understanding how they work and really pushing on them to and using them in more and more complex scenarios to investigate supply markets and supply risks and categories and, individual suppliers, right. And all these things as well as internal issues.

How do I develop a good procurement policy? Right. All these kind of good things just getting started. Right. And then folks will then get into the pilot phase. They start adopting some of the enterprise level tools, some of them from their traditional tech and service providers.

Sometimes they'll go more with start ups and, you know, testing in different test cases, they'll different use different platforms that are out there. You know, AWS is is a platform, but really starting to kind of get into the more sophisticated and advanced, you know, use cases. And as that happens, that's when things kind of get a little more complicated because it depends on who you are.

But, you know, generally you want to be working closely with it to make sure that whatever you're doing is, you know, let's just say there's full communication between procurement and it there should never be shadow procurement. There should never be shadow it. So having that linkage is really important.

[00:07:36 - 00:08:10]
Kelly:
Yeah, absolutely. And really flexing that leadership position of the procurement leaders is important. And, you know, we collaborate with purchasing organizations of all sizes. So anything from a solopreneur all the way up to, you know, very large multinational corporations. And so, you know, we really deal with lots of different challenges for lots of different businesses at different levels. And we're seeing everyone do, you know, kind of operate a little differently in this environment.

So, in your view, what are some of the major challenges that procurement leaders and teams are sort of facing today? And how can I help address these?

[00:08:10 - 00:10:22]
Pierre:
Sure. I mean, if I had to summarize, the challenges would probably be the military term, which is, you know, Vuca, you know, which is and now let me pick on the volatility and the complexity, the V and the C, aspect of it. And basically, you know, how are we dealing with the complexity that's happening out in the supply markets and with the rest of the business, just given what's happening in the, you know, geopolitical, you know, environment and tariffs and whatnot. But you know, so there's that kind of outside in problem as well as just how do we add more value to the business? How do we get more spend under management? How do we get broader business engagement and alignment?

How do we start integrating with other functions to broader areas outside of just core source to pay it? For example, around third party risk management and sustainability and, you know, etc.. So, there's a lot to do.

And in terms of doing more with less and AI is very helpful both on the, efficiency side. Right? So we can free ourselves up to do that through, increasingly autonomous kind of, you know, automation, better decision support, better data and analytics. And really letting us, do the things that are more strategic and more human, right?

And let the AI do other tasks. And I think the big thing is making sure to that the use of AI is a net add. Right. And that it's actually something that's not just some technology de jure, because it's not it is actually foundational. I've been a tech analyst covering procurement for over 25 years.

And, you know, this is not just, you know, something like blockchain, which, yeah, it's great. You know, transaction registries and all that. But it is a it is a fundamental, you know, sea change that is happening. But so it's tricky because I itself is a complex area. There's a lot of different types of AI and technology types and, you know, trying to manage that and reduce complexity, eliminate risk, you know, infosec risk, eliminate hallucinations, all these things costs.

There are some things you have to kind of navigate, which is what we spend a lot of time with, with clients doing. But, you know, but overall it is a net add and it's starting to really, pay some dividends.

[00:10:22 - 00:10:39]
Kelly:
Yeah. And thinking about procurement teams and sort of how things are evolving what type of new training or maybe new skill sets do procurement leaders need to be focused as we we get to the next stage of a lot of this technology?

[00:10:39 - 00:13:03]
Pierre:
Well, I think one of the biggest things is actually the people side of it. Right. And and so how do we kind of retool our skill sets, right. And how we actually, adopt the technology because quite often adoption has been one of the biggest challenges and why we continue to stay on spreadsheets and tools that are not necessarily, optimal. So, using I think this whole area on AI in terms of using just, let's say the chat bots and copilot is a is a good example, right?

In terms of how do we start to just get familiar with some of the basic tools and really push them to their limit and really demonstrate value and get some mastery over them, and then kind of take it to, you know, the the next level, right, in terms of being able to then add more and more, competencies.

So it starts kind of almost personally with personal responsibility and how we all choose to learn about this technology and get smart about it individually. You can take some online courses, you know, free online courses. There's just no shortage of content, but it's really pushing yourself personally. And then with your teams and your your peers within your organization and across organizations to learn what different people are doing within the organization, within it, within other functions, do some collaboration.

And then kind of like procurement should help, you know, stakeholders do look to what's happening in the supply markets. And the technology is over is really disrupting a lot of these, you know, supply markets and use that to kind of bring that value and pick the right tool and technology for the job to kind of drive more value.

And that means kind of leading by example. It means that procurement should actually be getting its own house in order. Using AI internally with its own operations. And you'll be and so that you'll be in a much better position to lead, and certainly to then just collaborate with other functions that are also trying to, you know, use,

AI in different ways, whether it's in kind of shared services groups or whether it's on the sell side or in court, you know, broader supply chain and finance, you know, these types of things.

So I think the the talent side and the organizational side of this, to really invest in people and for people to take, accountability and responsibility for their own development, to really learn this technology and apply it for their own benefit and personal benefit and job, you know, and job security, as well as to kind of help their stakeholders and the business.

[00:13:03:21 - 00:13:25]
Kelly:
Absolutely. I mean, I agree, there's so many different tools to use. And I think the information sharing among, professionals is really important. So, sharing what works, sharing best practices, developing new SOPs and things like that for new technologies, for procurement teams that are just kind of starting out that haven't really done much, have maybe heard about some of these trends in the industry, where should they begin?

[00:13:25 - 00:18:03]
Pierre:
The core playbook is still valid, right? In terms of procurement, spend management, supply management, you know, spend is what you pay supply; value is what you get. So it really still starts with stakeholders and spend and market and your strategy and how you create value and performance targets and capabilities and your value chains.

All of that still remains in place. However, the supply markets, first of all, are getting overhauled with AI having a huge impact for, for example, professional services and outsourcing and things like this are getting much more, digitized with, you know, a genetic technology and things like that. Starting simple and with your own, for example, use of some of the chat bots and copilots. And just being able to use them to be like, you know, a category spend category you've never addressed before. I knew, techniques, you know, that are best practice, right, to learn more about it and to ask about how other companies have done it and then start to really see how far the tools will take you.

[00:14:32 - 00:16:02]
Pierre:
So, for example, you can be like, all right, well, tell me, what's the best practice for a category strategy? And then you can say, well, great. You know, given the of what's most important there, develop a stakeholder question. You know, that I can use, you know. Bingo. Great. Now I want to do this for some spend category like, facilities management.

Okay. Tell me about the facilities manager market. Right. That tell me about how I want to come up with a category strategy that I developed before, apply that to facilities management, and then come up with a, you know, supplier engagement strategy for how I'm going to source and manage them based on my chosen strategy. It'll come up with that.

Right? So it's it's the the limitation is no longer becoming the technology. The limitation is really becoming ourselves. Right. And so starting with something like that and then being able to say, okay, now how do we apply some of this technology to build my own agent or to just start to use the tools better that have I for my existing technology and service suppliers?

And certainly telling them and being vocal about where you would like to have the I you know, kind of, help you. Right? Help them help you, little Jerry, Jerry Maguire there. Right. So, so I think it's, you know, that starting with, you know, the, the basics and then kind of, you know, getting into, as I was saying before, starting with, you know, stakeholder satisfaction or spend insights or level of self-service, level of automation, the amount of time that we spend wrangling data versus really analyzing opportunities.

[00:16:02 - 00:17:59]
Pierre:
So those types of metrics and opportunities, are still valid. It's just really now using the AI to help you in some of these, you know, key areas. So that might be around, you know, you follow the areas where there's like the most potential for, automation or touching stakeholders. So it could either be on the front end around, you know, intake management and, you know, kind of the front end of the engagement or with suppliers, supplier self-service, because suppliers are a huge, you know, set of users.

Right. So the more in which you can apply AI and CRM like functionality over to, you know, the suppliers that you're working with so that they can do self-service, they can check on the status of their invoices, you know, etc. and using AI to take their invoices in, auto extract, take, you know, take contracts, your own legacy contracts, see deep into them, pull out the key data and then be able to, act upon it, you know, and have the data be able to then integrate to your ERP systems, your source to pay systems.

These are kind of, you know, where the use case use cases vary, but it's kind of where you want to apply the AI, where it'll have the most impact. And that's going to be around complex data that you can kind of pull data out then that you can summarize. It's outside in data like market intelligence. It's kind of data that's in documents, not just invoices, but like in contracts pulling that key data out.

And then actually helping it, using AI and generative AI to help generate drafts, statements of work contract. You know what, you know, draft language, you know, things to kind of just help you and augment, you know, your work so that you can really elevate yourself up to the more, you know, strategic activities around what is the best strategy to for this, you know, spend category or, this supplier.

[00:17:59 - 00:18:03]
Pierre:
So, yes, the simple question, I kind of check it out a few different paths. But,I think those are some of the pathways.

[00:18:05 - 00:18:31]
Kelly:
Yeah, absolutely. I think any way we can really remove some of those repetitive tasks or tasks that are prone to error. Absolutely. I think all professionals would be on board with that. You know, in your research, have you found any mental models or framework to really help procurement leaders decide where to focus their efforts?

You talked a lot about metrics and really setting some goals behind those, those metrics, you know, what type of framework should, leaders be considering?

[00:18:31 - 00:20:07]
Pierre:
I think one thing you can do is actually start with the area around, a genetic framework. So once you've kind of mastered large language models and chat bots and, actually building an agent is a useful skill because within, in a genetic framework, what what are you telling an agent to do?

Right. And so you're telling an agent, here's a goal that you have in mind. Do you have to do a certain thing? Here is the context and the data that you're working with. Here are the tools that you might need to actually, you know, get this work done, you know, etc., and use it. And then here is kind of like the, the processes and maybe the other agents that you might want to interact with to get the work done.

So if you think about that, what you know, that's very similar to balance scorecards. Right. And so we need to know what are those goals. Right. So we should always you know, know what is the outcome we're looking for. Agile development is a great area. So, you know, work closely with it. But that ability to to develop things quickly and to build new capabilities quickly and to do that cross-functionally and to make incremental changes and to get adoption, and make sure that sticks versus just Big Bang giant, you know, mega mega projects is is valuable business process management are and how do we model processes is ultimately going to determine the types of agents that you'll need to do different things. And then like the whole area around operating models and operating models is really just like what's what ties people and tools to processes and metrics.

[00:20:07 - 00:21:14]
Pierre:
So, you know, as we start maybe pulling people out and putting an agent in or, you know, etc. for certain processes, it's really understanding how that is going to happen. So getting your as is your current state processes. And then as you start refining them and you start digitizing them, doing that kind of proactively and methodically, right. And as you do that, then you can kind of know the skills that are that you want to have humans possess versus what, so think about like Alexa, right, in kind of the Amazon world where if you get when you get Alexa, you can hook it up to a series of skills to operate different things.

And it's the same thing with kind of agents, right? You know, that skill might be turning up a purchase requisition into a P.O. Or getting confirmation back from a supplier or, you know, whatever it is. Right. So operating models and understanding the idea of that is an important framework because ultimately, you want to you want to evolve your operating model to become more digital, more self-service and more efficient and effective.

You know, ultimately, that's kind of what we're trying to do.

[00:21:14 - 00:21:35]
Kelly:
That's great. And you mentioned so many great metrics along the way to really track your success and to know if you're really focusing on the right things, which I think is great advice. You know, if we look ahead a few years, look at our crystal balls, if I adoption and procurement continues to grow and really pick up steam in all of these different areas, how do you see AI really transforming the procurement functionality?

[00:21:35 - 00:23:28]
Pierre:
Yeah. Well, I think one of the interesting things about procurement is that, you know, there's there's certainly a lot of runway to improve just within your core sourcing and supplier management and your purchase to pay operations. And doing that for all your segments of spend, your tail spend to your strategic spend, to your operational spend, you know, just doing that. But as procurement gets more evolved and what you see in the more sophisticated organizations is that procurement is kind of now stepping into enterprise level processes. So this area around how do we do third party management and third party risk management, third party onboarding, risk management from a supply chain standpoint, sourcing risks out in, not just my first tier suppliers, but second tier, third tier.

How do we model our carbon footprint out in those tiers and do some more advanced sustainability work? These types of things, it becomes procurement in the business units, and the functional partners, IT and finance, you know, etc. working together to kind of develop enterprise level capabilities. And and so for example, working on the customer side and making sure that we're, you know, supporting them and reacting faster, driving more revenue from them, doing early supplier collaboration to support product and service development.

You know, all these types of things are enterprise level mandates and imperatives where procurement should really be in the mix and be demonstrating competencies to be like, look, we already know how to do a lot of this orchestration work using AI for different spend categories. Let's take it a little further. And now let's say bring in, you know, the infosec and the sustainability things and really start turning this into a enterprise level capability around spend and supply management.

[00:23:28 - 00:24:03]
Pierre:
It's not only about, you know, externalization. Right. So a CPO should ultimately be a CEO or chief executive officer. How do we really take the power of these supply markets and harness them, especially with technology and AI, to bring them internally so that we can then bring more value to, you know, our customers? So I think that's probably going to be the biggest thing is kind of, expanding beyond the traditional kind of process, silos of core source to pay and actually being able to now, you know, getting to these broader, use cases to drive more enterprise value.

[00:24:03 - 00:24:11]
Kelly:
And this really is a great recap of your Spend Matters’ piece. I'm just wondering, as we close up, what's one piece of advice or takeaway that you'd like to share that you haven't hit on yet?

[00:24:11 - 00:26:39]
Pierre:
You know, I think I touched upon this, which was that biggest constraint is yourself and your thinking and not not the technology. And it's a challenge for procurement because we've always learned to about, you know, we need to kind of show that we, you know, kind of know everything in a way. Right. And you don't want to show weakness and and actually having the humility to understand and, and and having that, you know, knowledge that this market is around AI technology is like a market that impacts all other markets.

And it is something that is complicated. And you need to start just getting better educated on it. This is what I do for a living, and I struggle to keep up with all of the latest advancements. You know, what's what's happening in AI, namely on the agenda front in areas around knowledge graphs and ontology, you know, advanced semantic analysis, all this kind of stuff.

Right. That's happening. So we all have, some personal responsibility to address that knowledge constraint that we have so that because the more in which we learn and we learn and do, the less fear there is of the unknown, because we've already kind of we know what it is, and we've applied a lot of it, and we know what we want to do next.

Right? So you know that eliminating that fear and actually kind of running to the not the fire, but, you know, to the opportunity I think is is really the exciting, you know, piece of this. And the technology is shifting. I mean, it is, you know, because there's a lot broader enterprise things that are happening, you know, in terms of not just a genetic, but just the whole technology architectures themselves are moving more towards platforms, moving away from just throwing applications at the problem.

And now we have all this fragmented data. There's kind of new techniques that are kind of evolving. It is a wellspring of stuff to learn. So if you have mastered things around your core, you know, sourcing and procurement negotiations and change management and project management, you know, and, you know, Lean Six Sigma and Agile, that's great.

This is the next domain of knowledge that you must master because it's going to impact everything. So that is actually, good news. If you can actually go, you know, run to that fire and use it as a, as an opportunity to kind of, you know, build something good for yourself and for your company.

[00:26:39 - 00:26:59]
Kelly:
It's great advice. And again, I encourage everyone to go out and read the Spend Matters piece for more information.
I certainly learned a lot. And thank you for joining me and for sharing this research. And thank you all for tuning in to the Smart Business Buying Podcast.

If you enjoyed this episode and want to stay up to date, make sure you subscribe to wherever you tune in for your podcast. And if you want to learn more and read the full guide, visit amazonbusiness.com/podcast