Dr. Elouise Epstein is a Partner at Kearney with nearly 26 years at the firm, an author, and one of the most vocal critics of the procurement tech status quo. She built the spider maps that became the go-to reference for navigating procurement technology. She's now retired them. This episode covers why legacy SaaS categories are breaking under AI, what an AI-native procurement organization actually looks like, and why the old playbook of suites, portals, and conjoined vendor-consultant-analyst ecosystems is running out of road.
The spider maps are retired because the SaaS market is running out of gas. AI is obliterating legacy SaaS categories, and mapping vendors into static taxonomies no longer reflects how the market is evolving.
The "conjoined triangle" of tech vendors, systems integrators, and analyst firms created a self-reinforcing cycle of inflated ROI cases and failed implementations. Elouise reconstructed benefits cases after the fact and found them "comical."
Procurement 4.0 is outcome-based, not process-coded. You define the starting point, the end point, and the constraints. How the system gets there is not your problem. Process as we've known it is a workaround for broken organizational structure.
AI-native means building around an AI platform (Anthropic, OpenAI, Gemini) from day one, not bolting AI onto legacy data models and hard-coded business logic. A small team can ship an AI-native procurement app in two months. Legacy players can't.
For mid-market teams: stop buying transactional systems of record. Think about scaling capability through AI employees, not more software licenses. AI employees take KPIs and direction, do the work, escalate when needed, and don't care if the ERP dashboard takes 10 minutes to load.
Her first piece of advice for any leader: stop buying tech. Do one-year renewals. Don't make big, costly, multi-year commitments when the technology landscape is changing this fast.
Her second: get on the bus. AI and digital are where the bus is going. If you're in leadership, you need to be at the front of it, learning and driving adoption, regardless of your age or comfort level.
Governance in the AI era becomes a way of operating, not a policy document nobody reads. The question is what you allow, who can do what, and how wide the constraints are.
The people and culture gap is wider than the technology gap. The metric that matters most is failure: how often are you failing, how fast are you learning from it, and are you creating a culture where that's safe?
Guest: Dr. Elouise Epstein, Partner at Kearney. Background spans nearly 26 years in procurement technology strategy, from dot-com-era reverse auction startups to advising global enterprises on AI-native operating models. Author and creator of the now-retired procurement tech spider maps.
Andrew (00:06)
Welcome back to Leverage Supply Chains, the show where operators and builders turn AI and data into actual results on the shop floor and in the supplier network. I'm your host, Andrew Stroup And right now, if you're running supply chain or procurement, you're living in an AI solution zoo. Every vendor has an agent, a co-pilot, a suite, a portal, a marketplace.
Your inbox is full of AI promises, but your team is still living in email spreadsheets and an ERP that hasn't fundamentally changed how the work gets done. So today, we're going to cut through that noise and talk about what it really means to build supply chains for the AI era, not just bolt AI onto legacy tools and anything else you're being pitched. We'll use the retirement of one of the most well-known artifacts in procurement tech.
Eloise Epstein's spider maps as a signal that the old SaaS playbook is coming to an end. And we'll dig into what an AI native hacked together operating model looks like instead, especially for mid-market industrial distributors and manufacturers. To do that, I'm joined by Dr. Eloise Epstein, partner at Kerney, Arthur, and one of the sharpest and most honest critics of procurement technology and one suite to rule them all, thank you.
She's been calling out the hard truths for years and is now pushing the conversation into what comes next in the AI era. Elouise welcome to the show. For folks who might know your spider maps or your books, but don't know what you're focused on day to day right now, how do you describe what you do?
Elouise (01:39)
Actually, if we could figure that out, I would like to know myself. Actually, think Kearney would like to know as well. I'm not sure anybody knows. So I really look, I'm a provocateur. I look at what the future is and work backwards to today. And I'm not talking about
EGI and all this like robotic everything I but I look at three to five years out and look at what the trends are and how we can set this North Star and work backwards from that because almost everybody is stuck in the let's just call it the SAP morass or the Oracle morass or whatever. And it doesn't feel like we can get from here to.
AI future. so my job, at least my job is to break that down and chart that path from here to there.
Andrew (02:35)
Well, not an easy task, but I love it. And I love obviously thinking about, you what everyone called a futurist or forward thinking or setting some guide rails around these. mean, I think it's so interesting about this context gets developed over time and also the muscle memory around it that you get to see on that. You know, I think it's interesting because you have a very non-traditional path to your path of procurement and supply chains wasn't exactly
textbook, which I think there's generations of these people in supply chain procurement over time. But what were some of the key detours or left turns that shaped how you've seen the space today?
Elouise (03:09)
Well, my entire career is a left turn and the folks in my generation did not set out to do procurement. We ended up in it and that's true for me. And I came from the performing arts background. I was a production manager, not a performer, fortunately for everybody else. But it did give me good insight into skills that
certainly helped me today in terms of creativity and storytelling and so forth. And I was working at San Francisco Opera. It was a great job. I love the opera and I was doing a little work for San Francisco Ballet. Anyway, it's a long story, but it was, you know, and you're a San Franciscan. So it's like these two cultural touchstones.
And I thought I was gonna, I really wanna go from the opera to the ballet. And then the dot com run up happened. And I don't know if you were there in the late 90s, but I mean, you'd open up the paper, literally the paper and everything was all about this internet and worldwide web thing. And I just decided that I wanted to be part of it.
And so the first company I went to was, I made the leap from the opera to a startup that did reverse auctions and electronic RFPs back in 2000. So the technology was very rudimentary. I mean, to the point for the first couple of years, I would be on call. We'd have to reboot the server multiple times a day, just like bumps and bruises.
Andrew (04:27)
Hmm
Like
in the closet at the office, right?
Elouise (04:46)
Oh yeah, yeah, yeah, exactly.
exactly. was just like, fortunately, like within my first year, we did move to a legitimate data center, but oh my God, it was awful. Anyway, fast forward and that startup got acquired by, it was started by some Carney partners. It got acquired, reacquired by Carney or acquired by Carney. And then I've been there ever since. To answer your question, I got off in nostalgia there for a minute, but.
I crashed my career multiple times throughout my, because I just passed, I'm almost 26 years at Kearney. So along the way I've, I've made a lot of mistakes. And, but in 2016, I had absolutely just crashed my career. It was in the ditch. I lit the car on fire and I had to walk a hundred miles to my destination. And I was just standing around thinking, what happened? Like I didn't, I didn't, I couldn't even tell you. And
Andrew (05:22)
Hmm.
Elouise (05:39)
sort of make a long, long, long story short is that I came to, I came to realize at that time that there was, everybody had gone this way. Like for those that are not, that are just listening, everybody went all the way to the left. Definitely don't mean politically, but just everybody's over here on this one end of the spectrum and it's sweets, best debris, sorry, sweets, one thing. And what I realized is,
Andrew (05:59)
You
Elouise (06:06)
A lot of folks were dissatisfied with that. And at the same time you had Salesforce really flexing its muscle at the time in terms of its app store and its extensible
Andrew (06:15)
You
Elouise (06:16)
architecture. And I was looking at that and I thought, you know what? That's not going to be like, like, why do we not have this in procurement? So I went all the way to the other end of the spectrum and I planted a flag and said, this is where we're going. And, and
I was on my way to being right and people were starting to move across the spectrum to my side and then the pandemic hit and then everybody on the other side came all the way flooding to where I'm at. so then I, and now I, not only is everybody sort of where I'm at, sort of with the sort of SAS explosion, now we have the spectrum is extended further now to AI.
Andrew (06:59)
So crazy. Well, you I think your journey in the left turns to your point or whatever maybe crashing planes reminds me, you know, I'm an engineer, right? So I think about things in physics often. It's really, you kind of talk about that wave of the consensus, right? But yet the people here though, I think it requires like physics, like you need an outside force and perspective to shift that momentum and inertia. It's a very fascinating. And I think that's why I love, you know, about this job.
chatting with you and consistently over the years being able to, because I think we have perspectives that maybe, you know, a little, a little abnormal or against the beaten, the standard path. you know, the one thing actually, I think very fascinating about your background as well as like, you are one of the toughest critics in this space. You know, you've, you've literally written books, strong opinions published. don't shy away from calling out, you know, the stinky and what's broken.
Elouise (07:45)
You
Andrew (07:49)
Like what made you feel so compelled to take that public role and be so vocal? Like were there, what conversations did you want to force the industry to have that it was ignoring?
Elouise (07:58)
I, it's a great question because I don't think I've ever quite answered that way. One is I had nothing to lose. Like I said in 2016, 2017, even 2018, like I mean, I was gonna leave Kearney voluntarily or involuntary. I didn't have anything to lose. So by me just sort of like pulling the pin on the grenade and tossing it wherever it landed, that was, you know, that was a.
But like, what did I care? I did not care. And actually, at the time I was pre-transition, I was even angry. So like, if you saw me circa
Andrew (08:30)
you
Elouise (08:31)
2018, 2019, I was like, come on, come on, come get some, you know, like I wanted that fight. ⁓ I have since, you know, like moderated and become much calmer about the whole thing. But but the other piece, and this is the thing, this will drive me crazy is this
Andrew (08:33)
You
Elouise (08:51)
the conjoined triangles to quote Silicon Valley, ⁓ where you had the tech providers,
the systems integrators, the analysts slash consultants, and they're all scratching each other's back. And when I realized that, so it creates this group thing. So let's say, Andrew, gonna go start, you need to put a new procurement system. So what you would do is go to the analysts, the, let's,
you know, the Gartners, the IDCs, whoever, and they say, here's your top 10 players or top 20 players, you know, in a market where there were not legitimately 10, much less 20. And so you go there and then you go to the tech provider and they do all these benefits cases and like they're selling you and then they go to the systems integrator. And so they're all trying to get you because the systems integrator gets, you know, kickbacks from the tech company.
And then somewhere along the line, know, the consultants come in and sign off on like, if you invest $10 million, you're going to get $30 million in return. And what was sort of like dumb luck on my part is I came in after a whole wave, like 80 % of the market had done these sort of these implementations and they had all gotten to this some state of failure. And I'm not
Andrew (09:50)
and Hmm
Elouise (10:12)
castigating them. But there was enough evidence to say that that
was not a successful strategy. This conjoined triangle of everybody scratching each other's back and the collective group think of what a quote unquote best practice was. And so coming in after the fact, I was able to go back and look at and reconstruct some of these benefits cases after the fact. And by the way, will point out some of these were signed off by Kearney
Andrew (10:18)
You
Elouise (10:39)
So I'm not like throwing stones, like I'm throwing stones at my colleagues too, because that
Andrew (10:43)
You
Elouise (10:44)
was garbage consulting. Anyway, not to cut down that path, but the benefits cases, the ROI, it was just comical, or as the kids say, it was unserious. And so when I saw that, I'm like, first of all, I have to point this out. And second of all, you sort of add a little bit of,
Andrew (10:53)
You.
I love that. Well, maybe a little early in this new, you know, if we want to call it a super cycle, but it is interesting to think about how that backscratching is kind of emerging today in the lab.
Elouise (11:04)
and just like, what are we doing? So that really emboldened me to, and then you kind of get a little momentum on that and then it snowballs. And next thing you know, so I writing about it in a book and.
Andrew (11:32)
and the ecosystems and now the consultants and you know so TBD on what that means of the contrarian position of what that looks like and even the lens of procurement but it is interesting because I mean we'll get to this later too but I do think it's interesting where people should assume the A is going to solve all the problems but there is a foundational question set that I think sometimes gets lost over which is a whole nother set of questions but it does dovetail into your spider maps and so what I love about this is that
A lot of people, think, first discovered your work as it relates to the spider maps and the procurement tech landscape. I remember seeing them when they first came out and those maps literally became a public service for people trying to navigate the solutions. And so now that you've retired them, why did you decide to retire the spider maps? And what does that say about the moment we're in now?
Elouise (12:22)
Well, I think I sensed, so I'm not going to say I had some brilliant insight, but I sensed that the SaaS market was running out of gas. And now it's got a fancy term, the SaaSpocalypse or whatever they're calling it. it was, and so again, you have to think of this as an evolution. So this idea of one ERP or one source to pay didn't work. So we started adding more point solutions on top of that.
But that was never going to be the end game. We now have to move, you know, we're moving towards AI. And as you get to AI, what you find is much more customization and control. So if you think about it as you or I as a user, the AI is conforming itself to me, an audience of one. And I extrapolate that out to the enterprise is the
the solutions have to be more, and I'm gonna call it build versus buy. Of course, there's some nuance in that. Well, there's a lot of nuance. But what I'm saying is that if I'm pharma company one, I want it to work like I want it to work. And pharma company two wants it to work like they want it to work, but there's no commonality between, or very little commonality. And the way that we've done it before was, ⁓
Andrew (13:16)
Mm-hmm.
Elouise (13:42)
Well, Andrew, you and I are going to start a new company and we're going to have the one solution. We're going to adapt it to Pharma and we're going to adapt it to CBG. But at the end of the day, the core platform is the core platform. And I think what's really become clear is that I mean, there's a lot of other things like proliferation, cost integrations, blah, blah, like. But if you get past all that comes down to I want it custom for my enterprise because my enterprise is different.
than absolutely everybody else's. to really, and this would be my argument, is to really run a 21st century business, you have to have that core digital system conform to the organization and to the individual users.
Andrew (14:27)
Well, I absolutely agree. And I think it's very fascinating because I think there's that pendulum swing that occurs, right? That we think about how that operates. And I'm curious, because, you know, I think you touch a few good, really good points here, but your perspective on this, that like, you know, ultimately,
in this AI era, the legacy SaaS category models shift. I mean, in some cases, probably break is maybe one way look at it and the tools. And so I'm really, I'm really curious when you think about like, you know, the one giant suite and a portal for everything and like that mindset versus the AI wave and thinking through like how you think about monoliths, like where do you think we are in that pendulum swing cycle as we think about a different AI native world, if you will.
Elouise (14:51)
Yeah, 100 % break.
Well, think, so I'm glad you mentioned that. We are going to AI native applications and even let's just, let's overhype the term for a moment, bear with me, AI native or AI driven organizations. And I know that's a stupid consulting term and like,
But to me, what it's conveying is that the way we've been doing it in the past is not going to hold up in the future. Now, I want to make two branching topics from that. One, it's going to be painful. And you sort of alluded to the conjoined triangle thing that's going on with the AI chip makers and the LLM providers and love.
Andrew (16:10)
And.
Elouise (15:54)
So we're gonna go through a tumultuous period in infrastructure or in AI infrastructure platforms, blah, blah, blah. Just like the dot com run up, it pushed us forward and then it collapsed. But what came out of the collapse? The Web 2.0, the YouTube, Facebook, all of
social media, Amazon, even though it predated that, it's...
Like what came out of the collapse has shaped our world for the next 25 years or 20 some odd years. I think the same thing's gonna happen with AI. It may not be like an epic moment where it blows up, but it might have like its fluctuations. But there's no doubt that that's where we're going. And let's just say it all blows up here in 2026. What comes out of that blow up?
It's not like we're gonna go backwards to a single ERP system. mean, again, that's why it's
Andrew (16:50)
You
Elouise (16:51)
a spectrum of we did the suite, we did the point solution, now we're gonna do AI and we'll sort of evolve around that. And the other thing is, and this is what I'm working on in my next book, is I don't think, like this is the thing that I would argue we're missing.
We think about AI as automating the enterprise. It's not about automating the enterprise. To me, AI is going to expose the enterprise and expose its structural underpinnings. So when I say that AI native quote unquote organization or AI native, whatever you want to call it, it's this idea that what does an organization that leverages AI look like? Because a
Go back to my Pharma One example, whatever Pharma company or CPG company, like that organization has way too many people for an AI native work.
Andrew (17:44)
Yeah, absolutely. Well, you know, I think.
Jack Dorsey, you know, in the headlines in this moment today, like probably solidified a little bit of that position point in his recent release regarding block. But, you know, it's interesting, I would love to drill down, I guess, in that structure of, okay, so let's think about specifically like a mid market, you know, chief procurement officer or VP supply chain. And like they're in the stack today, they're hearing this podcast, they're thinking about, you know, they're seeing the sea of vendors and some say that they're, you know,
AI, but maybe they're pre architecture, or they're like AI on the deck, but they actually aren't maybe they are Anna, like, curious, your, you know, red flags, yellow flags, beige flags, whatever may be on how you think about that, especially as you think about where they should be focusing their time towards that future set.
Elouise (18:30)
So the first thing, first and foremost, I tell all my clients, just stop buying stuff. Just because, ⁓ or do one year renewals where you can get out of it easily. Just don't make big decisions and big costly investments right now because everything's changing. And by the way, things you, the way I look at it, the way you look at it, it will change in a year, in two years, in three years. So why would you sign up for a five year commitment?
Andrew (18:35)
you
Elouise (18:58)
with seven figures where the world could fundamentally change in many ways, ⁓ but technologically, mean, the speed at which AI is evolving and the capabilities, I mean, for me, my old human brain is not keeping up. So why sign up for this stuff? So that's number one, just don't buy lots of tech. Because one of my arguments is,
CLM is dead, why are we buying CLM systems? And yet, like eight out of 10 clients are putting in CLM systems, right?
Andrew (19:29)
So it's like literally the
one of the first things they want to try to install, right? It's like the simplest thing for them.
Elouise (19:32)
Yeah, exactly. it's
like, just stop buying that stuff. You're not going to be any worse off than you are today. Second, if you are a leader, it doesn't matter if you're CPO of Big Pharma Company or Little CPG or mid-market, whatever. Learn. You've got to learn this because most leaders, and I have stats to back this up, are over 40. So whether it's board level, CEO level, or even VP and director level,
pretty much over 40. And by the way, a lot of millennials are starting to get over across that 40 threshold. So like, you got to learn this. It's not native to us. It is much more native to Gen Z and definitely Gen Alpha, this AI thing. So we got to like understand our like our place in the world. And so if you're over 40, you got to be learning this stuff and working overtime.
because how can you make any decision about an AI platform? And I do nothing but study and work with this all the time. I'm not the right person to buy an AI platform because it's just beyond my sort of lived experience.
Andrew (20:39)
Yeah, absolutely. think, well, you know, being a millennial stepping into this 40s and being part of that, it is interesting because I think the vernacular and even the, you know, I grew up in the wave of like, let's talk about cell phones and what does the internet look
And what does the dial up seem like that step into a different, almost step change of accessibility. And I think we're seeing the same thing, even with like the newest wave right now, like gen beta, which is going to be fully AI data from the literally from the start. So it'd be very fascinating. ⁓ but actually dovetails into the other major topic I really wanted to chat with you about, which is this concept of procurement 4.0, right? And we're thinking about.
that or let's call it, know, AI native supply chains. You know, I don't think that we can easily just prescribe that to being co-pilots, right? And there's so much more to that language and so much more to like that scope. I would love from your perspective and as relatively plain language as it makes sense. So like, what does that look like? What does an AI native procurement organization structure in your head start to look like and form and operate? Because it's going to, I think inherently be different.
Elouise (21:49)
Yeah, undoubtedly. There's a lot to unpack in that question. So just raise your hand if I could go on into lecture mode. But this is the topic. So first off, I don't think anybody's in the Gen, sorry, Procurement 4.0 box yet. I think people are knocking on the door, but I don't want to be rushing to add another box in the evolution and then say, ⁓ well,
like, you know, everybody's AI native now. I mean, cause like, Koopa is now saying they're AI native. I will, I will, no, they're not there. It's a legacy architecture and the SAP is saying the same thing with their new, no, it's not AI native because AI native has a fundamentally different design construct to it. So, I mean, so, and I didn't want to, I don't want to be in the business of deciding who's in or out.
Andrew (22:23)
You
Elouise (22:42)
So what I want to do is give the framing and let everybody evaluate for themselves. I don't want to be the arbiter. I felt like in the spider map, I was the arbiter of innovation. But I didn't really want that. To do that, I really wanted to show the amount of innovation and the excitement that. I mean, I got a lot of excitement out of the innovation, and I still do. But I don't want to be the arbiter of it.
But the AI native, it's not just going to be a platform. So we can talk about the three that I think are the closest are LevelPath, Zip, and Oro. But there's other companies like Flipthrough and Rivio that have very specific point solutions or very specific productivity tools.
or agents in some cases that you can layer on into the way of operating. I would put them, like I would say they have a legitimate shot to step into that 4.0 box. But the broader, simpler way to think about this, and this gets like I'm poking at all the providers.
is that if you and I, Andrew, decided, you know what, we're done. I'm quitting Karni, you're quitting your, well, you already have done your own startup. But if you are a CPO, and this is sort of the little cheeky thing that I do, is like when I'm presenting to a leadership team and I tell them, let's just quit our jobs and all of us, we're gonna go build a Gen 4 procurement or procurement 4.0 startup.
Andrew (24:00)
You
Elouise (24:16)
what's the first thing we're gonna do? We're gonna figure out what AI platform, whether it's open AI or anthropic or Gemini. And then everything we do is gonna be built around that AI platform. So we're not hard coding like functionality and the coding we do do is obviously gonna be vibe coded. And so the five of us or the 10 of us in that meeting,
if we just immediately up and quit, we're gonna get to market in what, a month, two months tops? Whereas, take every SaaS software on the spider map and the legacy players, including you SAP and Coupa, they have a legacy infrastructure and legacy data models. And I guess I would say one more thing, I'm sorry, I kind of, I did.
get into my lecture mode. But this is not about embedding and coding the business logic into the application. This is really thinking, using AI means outcome based. I need a contract with this supplier, how the AI goes and gets it, and especially if it's working with the supplier's AI, I don't care. I only care about the outcome.
Andrew (25:20)
You You
Elouise (25:29)
Whereas in legacy tech, which is most of it, then what you're getting is you're getting business logic embedded, hard coded into the data models and the application layer.
Andrew (25:41)
It's really interesting because as you were describing that, I was thinking about like that classic consultant, you like you and I, well, you specifically, spend so much time in the consultant, I've spent so much time working with, or even, you know, being a tangential arm in some capacities as consultants, that classic people process tech diagram that happens. And you'd have to think, especially when it comes to procurement, like to your point, so many point solutions that took years to build and these, you know, coding the business logic into the data model and the work.
Elouise (25:58)
Yes.
Andrew (26:09)
close, that realistically the next whatever that four pointer looks like likely is a flip on that head where it's really not so much that it's actually focusing probably more on the people the process, right, which is okay.
Elouise (26:21)
But in an AI world, there is no process. It's outcome-based. don't care. This is the starting point. This is the end point and the constraints. And how you and the system figure it out, I don't care. The process is, I have a quote that I'm gonna use in a presentation next week. I can't find it, but it is like we... ⁓
Andrew (26:27)
That is true.
Elouise (26:45)
processes are making up for brokenness in the organizational structure and how information communicates through the organization. So that's why, like it's not people processing technology. And by the way, yes, there's a huge aspect to the people and the evolution, but we're sort of like the people processing technology is so out of date because
Andrew (26:54)
Hmm. you
Elouise (27:11)
What about AI employees in that equation? Because what we're going to become is humans, is supervisors. Again, we want outcome-based. I want you to go negotiate this contract and the AI employee can go do it. again, how it does it, I don't care. What I care about is the outcome. And so the process and people part of that, that whole construct is dead.
Andrew (27:36)
Yeah, it's very, it's very fasting the breach in which we're crossing in this category, right? Because like these fundamental tools, you know, you know, it's like these fundamental tools that consultants and others frame around are like literally reconstructing being reconstructed, which I think is the most probably fascinating part from like a anthropological like perspective. And as you zoom out,
I want to zoom in a little bit here because I think this is a good framing. It's like, okay, so, you know, the typical people listening to this pod are largely mid-market industrial distributors and manufacturers, right? So let's think like triple digit million in revenue, small teams, worked IT, a lot of emails, spreadsheet, you know, the gravity around that. What do you think AI native looks like at that scale? Like, what are they aiming for in the next, like,
to five years that you know they that they just they need to focus on otherwise they're gonna like miss the boat and whatever they're on.
Elouise (28:28)
So I think this is gonna be like bend the mind for a moment because I mean, I've given this a lot of like contemplation and sort of design because most of the folks in the mid market, not just the folks you're talking about, but even like the tech mid markets that are around your office and my office and.
is that you don't have a lot of resources. And maybe you have an ERP system. Obviously, the folks you're talking about have an ERP system of some variety. But a tech startup may or may not have a big tech infrastructure. And I think that's all OK. But this is not about, I think this is a step change in why I said stop buying.
Andrew (29:00)
you
Elouise (29:13)
technology and be more thoughtful about it is this is not about
buying more transactional systems of record. In this mid-market, what you want to do is expand your supply chain and procurement capability through the use of what I would say are agents, but actually within a year, AI employees. And so now you're sort of renting the employees or
think of it as AI temp labor, where they're coming in to do, source a bunch of events or to manage a category or whatever, because an AI employee, actually you give it KPIs, you give it direction and outcomes and it goes and does its thing and it escalates, it monitors, it tracks, it does things where it's like an agent's just doing a task or an agent to clusters.
series of tasks or of a workflow, but the employee's coming in to do this and the employee doesn't need, sorry, the AI employee doesn't need a massive infrastructure. And we're seeing this with OpenClaw, right? I just, which I'm gonna qualify this by saying, yes, I know this is a huge security risk, but if I just install it on my machine, now it's working as if it's Eloise. Now, again, huge security risk, but it can go into the SAP,
and it doesn't mind waiting the three minutes or the 10 minutes for the dashboard to load, it doesn't get irritated like I do. So, ⁓ yeah, or distracted. Yeah, so, so yeah, I'm highly distracted. Actually, I was late to our, just in full transparency, I was late to this recording, because I went down a rabbit hole anyway, about some research. Anyway.
Andrew (30:39)
Or distracted, right? Or, you know, all these things, right?
Elouise (30:56)
But yeah, so the AI employee just sits there and does whatever it does as if it were me. So it's not some like web bot scraper thing that's out there. It's literally just watch it. You know, it's just, it's behaving as if it was me. So for mid markets, why would you not scale out your procurement splicing capability? And why would you not shrink your finance, shrink your HR, shrink your IT? Because you don't need those people anymore.
And I think that's the big aha moment that we're coming to, but we haven't really broken through that concept yet.
Andrew (31:34)
Absolutely. you know, I don't have a full conclusive opinion yet on this, but very similar to you. I've been very much in the rabbit holes of all of this, it's OpenClaw or different Asian orchestration. It is interesting because I think the principle, the first principles of even engineering rigor still applies where as long as you, I think to your point, set them up with the appropriate scale permissions, whatever may be given the OKRs and KPIs and targets.
By structuring, I think the thing that most people actually miss in this category, by structuring that feedback loop and observability as a manager, then you can basically have them do whatever you want because to your point, you don't care about how they get there. You care about making sure they stay within the rails of getting to the end state, which I think is a hyper fascinating, but also requires a structured mind to then how to coerce it to do what you need to do along the way. And then you create playbooks around it and scale it thereafter, right?
Elouise (32:24)
So I might argue you need an unstructured mind for that, I think we'll see, and I'm not being a jerk about this, but I think the process driven that we've seen over the last, let's call it 50, 60, 70, 80 years is really an engineering mindset. Whereas I think when we get to AI, you can't control it in that way of hard,
Andrew (32:27)
Yeah. ⁓
Elouise (32:51)
coding decision trees. And I think that's going to be hard for people to wrap their heads around. And here's the real takeaway from, as we talk about AI and 4.0 technology, is that governance becomes a way of operating, not a policy document that nobody looks at. And so when I talk about constraints, like what is it that I will allow everybody to do and what is it I will only allow
leaders to do and so forth. Because those are going to be wide swaths of constraints, which is great. But not, like this is not, we have to be very clear on what we're governing. And oftentimes, I don't think we're anywhere near that.
Andrew (33:35)
Yeah, I, you know, there's some emergent new technologies, I think they're shaping some of that definition, what that is going to look like. But I think far from what I'd call concrete or maybe steady state there. Well, so you know, I think we've gone from like high level until then, like, okay, let's talk about, you know, who's listening to this pond. And now, I think, even more specific, what I have found to be true, especially when I go in and have helped other companies in the
Elouise (33:41)
Yes.
Andrew (34:02)
is that, let's zoom into the first 90 days, in that category of some new VP of supply chain, because I've often seen that's usually where they have the most influence of the change agent. saw this when I was on the White House. Basically, I told them, was like, look, I give you one good year, because that's the year that I can go in and be impactful before I become part of the board and the machine. I guess in that, whether it's in that type of distributor or manufacturer, what would you tell them to focus on?
in that first 90 days as they step into that role, which is going to find honestly love the lifeblood of a given manufacturer distributor.
Elouise (34:39)
Two really easy things to do. One is get a handle on all the technology and come up with a plan to reduce it, limit it. Because even if you're only paying, I don't know, $10,000, so it's like peanuts.
It's still consuming a lot of overhead and time and people and just the whole management of it. So figure out what you got, figure out what you can sunset and sort of what your AI adoption strategy is going to be. Obviously, adopting it today was almost March of 2026. Like if I came into that role in January of 2027, that's going to look a little different.
But I still
Andrew (35:22)
Uh-huh.
Elouise (35:23)
think you have to get ahead of this question. And then the second thing is, is I gather all the people in a room and tell them like this. is there's a bus going to the future. And if you want to be on the bus, it's going to be AI and digital. And you that I don't care what your age is. If you want on the bus, everybody gets a ticket. Now, if you don't want to get on the bus, that's OK. But you got to go and.
Andrew (35:28)
you you
Elouise (35:51)
And if you if I'm looking at my leadership team,
you guys got to be at the front of the bus like you can't be hiding in the back even if you don't know everything because that bus is going to the future and if you stand in front of it, it's going to run you over.
Andrew (36:06)
I love that because it was actually one of the thoughts that came up and I was actually had it written down was around like, it's, I categorize this as like a cultural or talent question, right? Like, fundamentally, like you step into that role to your point, the bus it's going, it's going whether you like it or not, there's no, you
Elouise (36:15)
It is, yeah.
Andrew (36:23)
the most helpful and also hurtful thing in life is time, right? And you can't really control it in any categories. And so the progression of that, is there any habit or ritual that you think or you observe that you think would be impactful for, you know, said leader giving him advice or her advice on how to handle that transition period?
Elouise (36:41)
Well, think so. It's funny you mentioned people and culture because that's what I talk about now is and this is also gets to one of the this was the byproduct of stopping the spider map is the technology just went through the roof and the market doesn't need me to talk about like to we don't need anybody to like plot the technology and do the analyst thing like that.
Just ask chat GPT, it'll tell you what you need to know. The issue is the people and the culture are way, way down here. So all of my work is with people and culture. And you get a spectrum. But this gets to, it's really about motivation and saying, look, I'm not gonna fire everybody. I don't believe in firing everybody because of AI. Now I know you and I, there's a lot of companies around us that are choosing that.
But it's not about that. It's about on one hand, it's enabling the younger generation that expects this kind of technology. And then those of us that are older is to help develop critical thinking skills, communication, relationship building, things that we grew up learning. Like it's merging those two. And I think as you come in as a new leader, you have to navigate that and figure that out. It is...
Andrew (37:50)
Mm.
Elouise (37:56)
I am not concerned about, like I don't, when I walk into big pharma company, I don't care what AI model they're using. My assumption is they're going to be using Anthropic or OpenAI or Gemini or something else. I don't even care because it's not about that. And by the way, like it's not co-pilot. Co-pilot, Microsoft co-pilot is just corporatized AI.
Andrew (38:13)
You
Elouise (38:23)
That's not what I'm talking
about here, right? But the question is, this idea of governance as a way of operating and how do I, because like there's a lot of productivity hacks that I can do. So how do I operationalize that within the boundaries of my people? And there are not good answers for that. And there are certainly no best practices. So if anybody's telling you the best practices, can.
you have my permission to kick them out of your office because they don't know what they're talking about.
Andrew (38:54)
No, I will take you up on that someday, I promise you. ⁓ Well, so that is interesting because I think that dev tells into our lightning round questions.
Elouise (38:58)
You
Andrew (39:07)
to field for you as we get to the back side of the show. In the lightning round, quick questions and your immediate reaction responses, which I know you're quick on your feet on, if you had to pick one metric for like a CPO or VP supply chain to track, whether their AI bets are actually changing how work gets done, not just adding more dashboards, what would that metric be?
Elouise (39:28)
Failure. How many times are you failing or how are you going about correcting a failure? not to, I know this is lightning round, but we talk about fail fast. What are we really talking about? We're talking about learning quickly and because humans learn by failure. And so often I've seen this for like 20 years in the whole digital, the march to digital. And now I'm seeing it again with AI. We become stuck.
Andrew (39:36)
Mm.
Elouise (39:56)
and we don't wanna start because we don't wanna fail. Like failures, like my greatest learning moments have been because of failure. Like I said, I crashed my career into the ditch multiple times, but by doing that, that's the only way that I'm still with Carney is because I learned from that. And so you gotta have a metric around failing. that doesn't mean you wanna crash your company by the way, but I can't just wait till it's perfect before I will stick my toe and say, hmm.
Okay, maybe.
Andrew (40:25)
Super interesting. I know it's also supposed to be a lightning round, you know, think, you know, we have these great conversations anyways, but it is interesting, right? Because you have obviously reversible versus irreversible.
Elouise (40:34)
Yeah. Yeah.
Andrew (40:35)
But to your point, I think the failure feedback loop can be faster, but it is going to feel different for people, right? That experience, which is different than like, you know, pulling up a tool that doesn't click the way you want it to. It's very different than basically being told either the output is wrong and or you're wrong. And I think it's a very specific visceral difference, right?
Elouise (40:53)
Well, and also most, at least in corporations, failure is seen as a bad thing and you're punished for it. So you do everything, including fudge the books or create excess, whatever, and it doesn't have to be like, devious. It's just like, we kind of naturally do that because we don't want to be called out and we don't want to be punished. Look, I do that too.
But the fact is, that it sort of stunts your learning. And when I talk about AI exposing organizations, that's gonna be one of the things it does is the people that take extra time and build extra processes to prevent being seen as a failure or being seen as having failed, like AI is gonna cut right through that.
Andrew (41:39)
Absolutely. Next question. Finish this sentence for me. If you're still treating procurement as a back office function.
Elouise (41:46)
You're operating in the late 1990s.
Andrew (41:49)
All right, well, as you know, sometimes industries trail in this category, so we've got to help them along the way. Okay, overrated versus underrated. Go ahead, sorry, Eloise.
Elouise (41:57)
All
I want to say though on that is that there's like, if you think about a company, just go down to the basics things, there's revenue and there's cost and cost goes to employees and cost goes to suppliers to produce the product, ship the product, whatever. that equation is very simple. It's a two part equation or 2.5 part equation.
Andrew (42:04)
Yeah.
Elouise (42:22)
how can you not, like everybody focuses on revenue and growth, but if you're not operating a profitable business, what are you doing? And you and I like look around, like open your window and look at the tech companies around you. They've made this mistake. The older ones have made this mistake repeatedly. anyway, just want to, like, so when we talk about procurements back off, like it can't be like just the math doesn't math.
Andrew (42:31)
Hahaha.
Mm-hmm.
100%. I think that's, you know, I think I think I got exposed a little bit during COVID. And I think it's getting even more exposed now in the scenario of a whole bunch of different convergences of motions and movements happening in AI. Okay, last question before I get to the fun one. It's if if someone wants to see through like, you know, the AI hype and kind of drill into more of what you're talking about when it comes to the layer of supply chain procurement, is there like a book
paper or an operator you'd want to point them towards, besides your upcoming book, by the way, which I'm very excited about, but anything that comes to mind that you think, start going, like start digging here.
Elouise (43:23)
I have a very simple answer to that. So first off, you want to learn about, so my push is everybody needs to learn about and do AI. So if you want to learn about and do AI, please, please, please stay off of LinkedIn because the people posting about AI and LinkedIn have no idea what they're talking about. And it's just trash. So,
Andrew (43:37)
Thank
Elouise (43:43)
Sorry, I'm getting excited. If you want to learn about AI and C with coming, you got to follow creators on YouTube or TikTok because that's where you'll see people doing this and they'll deconstruct within a couple of hours of every new announcement, Gemini 3.1, know, cloud code, this, that, the other thing.
Within hours, you'll have a hundred videos of people breaking down the key changes or hacks to make Notebook LM better. I'm a huge Notebook fan or whatever it is. And they're doing it. All of these things are doing, let me show you. me, like one of the things I was, last night I was trying to enhance a slide deck. So I put it into Google slides.
and wanted it to clean up my deck and do all the formatting and it didn't work. And so then I went to YouTube to figure out how to do this. And I went to Gemini as well. So this is what everybody needs to be doing. Now, I'm not giving you too much more prescription other than that because maybe you only use Anthropic or Clot or maybe you only use Palantir or whatever.
I'm not trying to be like, we all are down different sort of rabbit holes and that's our interests. So follow that. But I guarantee you there's gonna be a thousand videos on whatever AI you're learning.
Andrew (45:07)
Absolutely. Yeah, the distribution is wild at this point, which is, think, an interesting accelerant in many ways. Okay, well, Matt, last major question besides, you know, some like, just logistical ones. And I love this question, by the way, it's if you could go back and give a younger version of yourself, you know, or just somewhat entering procurement supply chain today, one piece of advice about how to think differently in this space, what would it be?
Elouise (45:31)
You have to believe in yourself because nobody is going to believe in you and nobody's going to promote you, know, like, you know, like sort of like talk you up, hype you up. I kind of walked in thinking, you know, like, I, you know, like, I'm here, I've arrived. I also walked in and early in my career thought I'm here to learn. Like, I want to learn from all these smart people around me.
Well, it turns out that they weren't any smarter than I am and they didn't have any more experience than I did. And in fact, in many cases, technologically, I was way more advanced than them. And if you would have told me that before I left the opera, I'm like, there's no way, there's no way. I'm like, I'm the dumbest, I'm the slowest, I'm the least technical. And I was surprised and I have been surprised.
that plays through in terms of business, in terms of technology. Like nobody knows it all. And I got psyched out for years of like, you know, there's all this jargon and people are throwing around, you know, like EBIT and EBITDA. Like I'm just like thinking, like I'm like Googling in the background, like I don't understand what this stuff is. And I think you just have to have the confidence in yourself because nobody has it, nobody else figured, look,
Look around the room, nobody else knows. Here's a great example, I know I've extended this question, but, and especially in consulting, there's always a lot of charts and graphs and stuff. And I just look at the charts and graphs and like, I can't read them. Like I just, like my brain doesn't process the information. And people are like, know, go deep, you know, go into that chart, go into that bar or that, you know, like chart of the.
part of that, you know, what's this and that, and it's just like, goes right over my head. It's just because I have a different way of learning and interpreting and, or rigid processes just like turn me off. Like this just like, and, and, and, because I'm not an engineer and I have this unconstrained brain, but I looked at that as a liability for years and years and years. And the way I would reframe it,
for younger folks is we all have our spikes and we all have things we're good at and the things we're bad at don't focus on them because it's not going to help. If you want to improve them, great, but if you don't, that's okay too. But just know what your skills are or your superpowers.
Andrew (47:57)
Love that. Superpowers are important and they're not always obvious on the onset so I think it's I really like that as well.
Elouise (48:02)
And they're definitely not the same, because
I'm sure you have different superpowers than I do.
Andrew (48:07)
Yeah, I'd imagine so. yeah, it's super great advice to give, especially thinking about in this transition point that we're living through right now. Well, Eloise, this has been fantastic. Obviously, thank you for walking us through how to move beyond suites and portals, how to retire the old mental models like the spider maps, and what an AI native hacked to get a supply chain might look like in the real world.
For the listeners, if this conversation helped you see through some of the AI hype or maybe gave you a career 90-day starting point, share this episode with your procurement operations and finance teams. Follow Leverage Supply Chains wherever you listen. Leave a quick rating or review if you can. It helps other operators and builders find the show. I'm Andrew Strupp. Thank you for listening, and I'll see you next time.