Tennessee on Supply Chain Management
Go beyond the headlines with co-hosts Ted Stank and Tom Goldsby as they tackle the trends, challenges, and big ideas shaping end-to-end supply chain management. Each episode brings practical insights, industry perspectives, and lively conversation to the issues that matter most to today’s supply chain professionals.
Enjoying the show? Download and subscribe to Tennessee on Supply Chain Management wherever you get your podcasts.
Tennessee on Supply Chain Management
S5E1: Beyond the AI Hype: Auger's Dave Clark on Building Smarter, Faster Supply Chains
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In the Season 5 premiere, co-hosts Ted Stank and Tom Goldsby welcome Dave Clark, founder and CEO of Auger, former CEO of Amazon’s worldwide consumer business, and a 1999 graduate of the University of Tennessee MBA program.
Clark explains why AI’s biggest supply chain opportunity is not simply automating tasks, but connecting fragmented systems and shortening the time between signal and response. He describes the “coordination tax” created by disconnected data, meetings, and competing incentives, and how AI can shift organizations from optimizing individual functions to managing enterprise-wide outcomes.
Drawing on 23 years at Amazon, Clark discusses why companies over-invest in forecasting and under-invest in responsiveness. He also considers AI’s impact on the workforce, the enduring need for leadership and frontline expertise, and the importance of redesigning processes around customer service, margin, and working capital.
Along the way, Clark shares lessons on delegation, information-security risks, personal AI tools, and why readiness for transformation is more mental than technical.
The episode was recorded live at UT's Fall Supply Chain Forum in Knoxville, Tennessee, on November 11, 2026.
Related links:
- Learn more about Dave Clark and Auger
- Listen to our previous episode with Clark from July 2025
- Save the date for the Spring Supply Chain Forum, April 6–8
- Download free white papers from UT experts, including our newest, "The Readiness Gap," produced in collaboration with Auger
- Become a GSCI partner to learn, network, and recruit with the top supply chain education institution in North America
- Join the Advanced Supply Chain Collaborative to explore advanced concepts in SCM with top industry experts and scholars
- Follow GSCI on LinkedIn
- Subscribe to GSCI’s monthly newsletter
- Read the latest news and insights from GSCI
Live From The Supply Chain Forum
Intro & OutroWelcome to Tennessee on Supply Chain Management, where we unpack the drivers shaping today's global supply chains. From innovative ideas to real-world solutions, each episode brings you insights from the leaders charting the course for our discipline. Now here are your hosts.
Ted StankIt's a great pleasure to be here with you this morning. I'm Ted Stank, one of the co-hosts of the podcast. This is season five, episode one. We've been doing this for four years now and are in our fifth year. Tom, welcome.
Tom GoldsbyHey, great to be here. I think uh maybe for the benefit of our listening audience, we should kind of set the scene of where we are. Okay. You know, we're capping off the three-day festival of supply chain that is the supply chain form here. Rocky Top before us. What is it, like uh wait, what's the estimation about? 15,000 people. 15,000.
Dave Clark15,000 people. We had the
Tom Goldsbywave going earlier. It was just amazing. Crowd surfing. We got the mosh pit up front. It's it's really a great scene. It's a little
Ted Stankbit chilly here in uh Nealand Stadium, though.
Tom GoldsbyIt is. It is. But we're gonna persevere. It's been a wonderful past couple of days. We had some really uh great networking on Tuesday afternoon and then uh a full day of informative sessions yesterday. And hey, it's the podcast this morning. What a great way to get day three started.
Ted StankAnd what's even more exciting is starting off our fifth season is our guest. Dave Clark has been a long, long time supporter of our program and on our initiatives. Dave, I think this is your second time with us on the podcast. And actually, you were responsible for us changing our format a bit. We were used to this banter between Tom and I about what was going on in supply chain current events, and then we would go to the guest, and you were like, wait, I got things to say about this too. So we're like, yeah, so now we don't do that little bit. No, I was
Dave Clarksupposed to sit in silence until you guys talked about all that stuff. I was like, go.
Ted StankFor those of you who don't know, Dave is uh 1999 graduate of our MBA program, an Auburn undergrad. So he just traded one color orange for another.
Tom GoldsbyMusic education undergrad.
Ted StankYou were our junior high school music teacher. So music is a big part of supply chain management, I think. So yeah.
Dave ClarkHey, if you can teach seventh graders, you can manage again.
Ted StankAs Tom mentioned, Dave worked his way up through the ranks at Amazon to become CEO of Amazon.com and left Amazon and is now the founder and CEO of Augur. And so one of the things we were gonna try to unpack today is his vision of the future and how Augur plays a role in that and broader supply chain issues as well. So let's
Dave Clark’s Practical AI View
Ted Stanklaunch into this. Dave, what is your vision of what supply chain is today and where it goes in the future? There's so much conversation about AI and the role AI is gonna play, and you know, we're all gonna sit here in 2030 and push a button and everything's gonna happen.
Dave ClarkWell, and you know, we're supposed to be in autonomous cars like three years ago. Yeah, exactly.
Ted StankSo
Dave Clarkit's I mean, I have a fairly pragmatic view, I guess. I'm a very optimistic AI person, but also been in supply chain long enough to know that no matter how this thing plays out, you're still gonna have lots of incredible people in supply chain. If I think about like the last 30 or 40 years, there really hasn't been a lot of like transformational innovation, right? Like, I mean, you're talking like barcodes and shipping contain like you
Ted Stankknow,
Dave Clarklike what's the most of the technology support in supply chain has been all about vertical optimization and you know, and a lot of that vertical optimization has been kind of wrung out, right? Whether it's WMS or TMS or what have you. And the big gap has always been how do you connect these things? Like how do you actually move from managing supply chain in siloed verticals into managing into horizontal outcomes? And even at Amazon with you know, tens of thousands of software developers, I still couldn't do that, right? Like that wasn't a thing you could achieve because you could automate like crazy within the stack. But you know, when you wanted to bring things together, you still had this issue of you gotta have people in rooms and meetings with spreadsheets who are bringing their truth out of their system together with another group. Think about the holiday, it's like two or three billion dollars a day. You have thirty days to do it, sixteen days matter. So you kind of can't do a weekly planning cycle in that. Every day is like a fortune thousands annual revenue, so you can't miss a day. So we did what all large technology companies would do. We built the mother of all spreadsheets and integrated everything. It's the closest I've ever been to actually having supply chain on a single pane of glass. Technically, not possible pre-AI to really make that stack work. But now you can't. And so that's where I think it's going. I think the real move is finally us being able to move to a leadership world where we can manage around outcomes the way you know people have been talking about wanting to do in pods and every other format for you know decades. Yeah. And being able to center around customers and value, you know, the actual margin improvement, actual customer service improvement, actual new product introduction speed, as opposed to what was my transportation cost today.
Ted StankAnd you know, as I look back on it through my career, I've been doing this for 36 years, and I wasn't here for containers and barcodes, but I was here for QR codes and RFID and blockchain. And it's almost like it's
Dave Clarklike an incremental barcode.
Ted StankIt's yeah, it's almost like we've been building to the point where none of it was going to really optimize and reach what we needed to until something was able to come along intelligently and pull it together. Russell Algore was talking yesterday about uh the need to get inside of decision makers' minds. Because I think on a daily basis, each of us make thousands of decisions based on inputs that we receive without really thinking about how do we make those decisions. We've been making them, we've been learning about it since we were infants, probably, and now we just make those decisions. And in a business world, if you talk to your planners, if you talk to your buyers, if you talk to your uh decision makers at various points along the supply chain, most of them are making those decisions without even noticing the decision process. And so we have to dig into that to understand how to take this information, put it together, and make better decisions.
Dave ClarkYeah, I think it's a part of the piece that gets lost in the AI conversation is that you know people keep talking about AGI or you know, generalized intelligence or you know, superintelligence or whatever. But really to run a business on AI, you need a couple of other pieces that you need auditability, like you need the ability to actually, in the same way you would ask a person why they made a decision, you need to be able to ask an agent how it made a decision. And it needs context. And it's just another way to say what you said, which is like all of the things that are in everybody's head, that are in the 10,000 macros that run the Excel's that run the business, that are in all those process documents. It's how do you embed that into the data structure so that agents have access to it and they can operate it. But I think the most interesting thing, the thing people we miss out in the discussion is we spend all our time talking about all the decisions that are made every day and that you know, can AIM make those or not? And we don't spend enough time talking about the decisions we don't make every day. Supply chain is a tough business and it's contact sport, but it is too hard to just get a shipment to a customer, you know, or get the POs in the right place or the inventory in the right place. And because we spend all our human capital on that resource, we don't actually get to the process improvement work. So I mean my thesis on this is that we're sort of asking the wrong question in some ways, which is like, let's not ask, can the AI like allocate shipments? How do we make it allocate shipments so I can actually go do the stuff that I really wanted to do when I got into supply chain?
Ted StankThat's the really exciting thing. You talk to any manager about have you really had a chance to sit back and say, how could I really optimize what we do and make it better? And the answer is yeah, I think about it all the time, but I'm so busy putting out fires that we never have a chance.
Dave ClarkOr and you can't actually see it. So to solve a problem, you have to see the problem. And you kind of have to let the tide go out a little bit, clear a bunch of stuff, and then see what the next set of rocks are there. And you know, I think that's you know sort of missing dialogue here is how much of the sort of continuous everybody knows one problem begets another, right? Like you solve one problem in supply chain, there's another one waiting. I think that's an infinite loop. At least I I've never seen the bottom of that pool. So, you know, every time we pull the rock back, like you're gonna find another one. And so there's no shortage of things for people to do. It's just they're gonna be harder, more interesting problems.
Ted StankSo that's really the more
Dave Clarkvaluable problems, too.
Ted StankI love that metaphor about the tide going out so you can see those rocks.
Tom GoldsbyIt just doesn't feel like it's gonna recede anytime soon. Right. Hey, last night at the reception, I pocketed uh a cushy that uh said, you know, dealing with the uh coordination tax. Yeah. Can you tell us a little bit about the coordination tax and kind of some of the pains by virtue of having decision rights distributed throughout the organization, not seeing the problem? What are the real costs to organization?
Dave ClarkCan you raise your hand if you like meetings? Just for everybody on the podcast, there were zero out of 15,000 hands raised.
Ted StankRight.
From Silos To Outcomes
Ted StankNot
Tom Goldsbya one.
Dave ClarkBecause meetings suck, right? Like most meetings are not that useful. Most meetings are like coordination tax. They are a place where people from different teams go to the room with their data. People argue over what truth is, people negotiate for their own incentives, all in positive intent most of the time, but like you know, people have their business to run. Nobody's running the business, everybody's running their slice of the business. And usually what comes out is consensus, and usually that comes out after the best answer has already passed. And so it just mostly is not a fun way to live. Like you'd much rather be like solving problems and doing real work instead of trying to negotiate on whose numbers right. And I find that easily in my past life and in almost every customer we talk to, 60-70% of the meeting time is actually on this coordination work. Basically, the systems that run supply chain have turned us into human APIs. Our job is to shuffle data between systems in the supply chain via other people. We're
Tom Goldsbynot very good at it.
Dave ClarkWell, we have a a lot of loss. We're high latency in signal loss. And so and it's also no fun. I don't know that many people who love that work.
Ted StankAnd you're sitting in a meeting realizing there's all those fires starting while you're in that meeting, too.
Dave ClarkWell, you're also fighting the fire that started four or five days ago. Right. Meanwhile, another one took a while to get more started.
Tom GoldsbyFind a meeting room, make it happen.
Dave ClarkJoe wasn't available. He sent a delegate. That delegate doesn't know s**t about what we're doing right now. They can't decide. Joe finds out you got to have the meeting again because Joe is like, what did you agree to? I'm not doing this. We've all had this, right?
Ted StankHow are you finding senior leaders reacting to the message that you guys try to try to provide?
Tom GoldsbyAnd to whom are you directing the message as well? We talked about that a bit last time.
Dave ClarkI think this transition really has to largely be led from the top. It's sort of tops down, bottoms up, which to say that like the vision has to be clear from the top. And the best people we work with often don't necessarily know the how, but they know the where. They're not sure exactly the straight line between here and a vision of a more connected autonomous supply chain, but they know that that's the destination and they're starting. It's harder than it seems. And then once you're starting, then it's about engaging people who do the actual work because that's where the context lives. That's the whole magic of making this agentics work is to put the context into the structure of the system. So the best people doing this are just committed to the journey, I would say.
Ted StankDo you find that there's a lot of pushback from C-suites that don't necessarily grasp the notion of supply chain and the way we've evolved over the last several decades?
Dave ClarkI don't know that there's pushback per se. Like I actually think there's a tremendous amount of tops down pressure from boards and C-suites to do something. Like, you know, sort of like do something. You should save a lot of money. You know, and we see a lot of people getting sort of go gets, you know, I need you to go get X amount with AI or with something else, or you you can't have more than X amount of headcount, you need to figure out automation or autonomy. So we see a lot of that. Uh what I see more than anything is disbelief. A lot of people have tried AI sometime six months ago, and it gave them a stupid answer to some question they asked, and they were like, oh, this is just another one of those things. Or they've been in supply chain a really long time and they've been promised some version of this 6300 times and not gotten it. And so there's some maybe rightful slash cynicism slash, you know, disbelief in the stack. Yeah.
Tom GoldsbyYeah, an awful lot of that ready, shoot, aim mentality. Just go get it. Yeah, without a real vision. Right, yeah. Without any vision or sense for what they want.
Dave ClarkYeah.
Tom GoldsbyYeah. So uh as you talk about a connected, autonomous supply chain. We've got a lot of students in the room today that are thinking about their futures and entering the realm of supply chain. What does that future look like for the future workers in that supply chain? What work is left behind, what decisions still need to be made? What roles do they need to prepare them to move into?
Dave ClarkI think it's great, actually. I think it's a future is very bright for people in supply chain because, again, there's a limitless stream of problems, and many of those problems are not solvable by AI. Uh we had a great conversation in that breakout yesterday with the uh with Kevin and the Shaw team, right? They were talking about you know going out and getting frontline leadership experience in plants and warehouses. Like I don't care what's happening in robots, there's always gonna be a need to lead people and put people out in those environments. I mean, Amazon has probably more robots deployed than anybody else outside of China, and I think they're still hiring like 300,000 people for this holiday, right? So it's the need to have a skill to lead people is not going away. Maybe the good news, bad news. The problems you're gonna face are not the simpler problems of moving things through a system. They're gonna be complex. How do you solve a network? How do you fix the supplier? How do you find a new supplier? Supply chain's a series of contractual and relational connections, right? And so connecting systems and processes is is one layer that opens it up, but it opens you up to bigger problems. And uh which is great. I mean, I think that's why most people get into it. But supply chain is gonna be harder, it's gonna have to move faster, it's gonna be more strategic within the companies themselves, and it's gonna be much more technical uh than it is today.
Ted StankDave, you and I have kind of shared this optimistic vision of what AI and the future is gonna be of work. And you know, if I'm a student sitting here, I would far rather tackle those big problems than feel like my job's gonna be going in and staring at spreadsheets every day and you know, figure out how to move this this box from here to there.
Dave ClarkI would trade the office for a shop floor every day, all day.
Ted StankYeah.
Dave ClarkBy the way, I will tell you one thing that is a hundred percent true there is not a single person at any of the foundation models that have ever made or moved anything in their f**king life. And if they had, they would not make the claims they make on AGI. Because if you've lived in the reality of supply chain where the world basically wakes up every day trying to mess up your day in every possible way, from politics to weather to employees to whatever, the the stream is too big. Future's bright in supply chain still.
Tom GoldsbyI think it would be uh a big miss on our part if we didn't talk about your 23 years at Amazon and how it's kind of led you to this state of mind.
The Coordination Tax Of Meetings
Tom GoldsbyAnd I mean, I I think all of us in the room will recognize, I think the term was brought up yesterday that we've become diabolical customers, and it's largely thanks to you and the organization that built up.
Ted StankTell the story about your son when he was little.
Dave ClarkOh yeah. My son does not agree with you on that. Oh no? No, no, no. Okay, let's see. My youngest son, we were coming home from somewhere one day and he wanted something. I don't remember what it's a Lego or something. He was like, I was like, okay, I'll order it when I get home. He's like, where are you going to order it from? I said, where do you think I'm gonna order it from? I'm gonna order it from Amazon. He's like, no, I want it fast. I've never actually pulled the car over and done one of my parents did when I was a child, but I almost did. It's
Ted Stanka look over the seat, and it was like every pair of those, right? It
Dave Clarkwas a silent long pause after that.
Tom GoldsbyBut I was sharing with you.
Ted StankSo those diabolical customers are now maturing.
Dave ClarkIt is always the case. Customers are divinely discontent. Customers will never want less stuff. They will never want to pay more, they will never want it slower than they get it today. You know, everything you do with customers, every notch in the ratchet just earns you criticism on that notch in the ratchet so that you go to the next one. That's the game.
Tom GoldsbyYeah. But I I was sharing with Russell that I think I may have been one of the first 10,000 Amazon customers, you know, back when all Amazon sold were books. Ted, I think you might have had a few hairs on top of your head back then when we were first meeting each other.
Ted StankPossibly. Possibly.
Tom GoldsbyAnd then of course you went to A to Z, but it you know, seven to ten days was the standard service. And then there was like, I don't know who got the idea, maybe it was Bezos, two days. Two days. And then there got to be the race to get sub-two days. And just the nature of that build-out is exorbitant. I mean, you can deliver nationwide pretty much. I mean, North Dakota might go lacking if you have four distribution centers, right? But to get sub-two day, next day, same day is just a different level.
Dave ClarkYeah, I mean, a lot of that just has to do with scaling over time. Getting from you know, a week to two days is deep process discipline, right? It's it's just getting into like taking buffers out of everything else, right? It's like how do you do improve decision velocity, right? It's uh compressing that time cycle down. Getting from two days to one day and same day is physical proximity, right? It's about owning each piece of the pie so that you can control exact timing of when things happen, right? Because you need to shrink the time between signal to action on each step.
Ted StankTime space.
Dave ClarkUh and it needs to be closer. And that comes over time. And so two-day was a process evolution more than a network evolution. One day, same day is more of a network proximal evolution than it is, you know, a process evolution, if you will.
Tom GoldsbyAnd then do you believe that we've kind of taken that about as far as we can short of distributing through store, physical store?
Dave ClarkYeah, I I mean, I think I think sort of same day, two to three hours is kind of, you know, you know, or within an hour if you're like super look at this, probably sort of a bit of a question of what's the the the sort of diminishing returns at some point in that. The bigger question is selection. You can have a a piece of selection that's available at this speed and the whole pie of selection available at another, and that's always been the the interesting balance for a lot of these businesses. It's like what's the great
Ted Stankrecent story. Friend of mine was at my house in Beaufort, North Carolina, lives in Richmond, took off on a Sunday morning, calls me an hour later, said, I forgot my computer charger. Can you mail it to me? So it's like, yeah, I can do that. And in the UPS store in Boefora, you could have bought four of them by the price. So I'm getting ready to go and I'm like, wait, this is stupid. I got on Amazon. She lives in Richmond. There's a DC in Richmond. I got on Amazon, I ordered it. By the time she got home to Richmond, it was on her doorstep.
Dave ClarkTo ship it there, probably you could have bought three of them in Richmond. Yeah, yeah, exactly.
Ted StankRight. So and you know, so again, you talk about what goes on in our minds and decision processes. You know, I'm of an age that if I had to get something from here to there, I better go to some package delivery place, right? And it's like But
Dave ClarkI do think this is like the takeaway I think for most companies should be people get a little over-indexed at times on the forecasting and not enough index on decision compression. You know, it's sort of this being faster is almost equivalent to predicting better.
Ted StankExactly. Yeah. Yeah. We've been doing a lot of work with planning, and uh Mike Burnett's here, and uh you know, we we're talking that if you can get fast on responding to demand, you don't have to be as good about predicting demand.
Dave ClarkYeah. I think one of these things that is just fundamentally flawed in the way the world has thought about supply chain a bit is this idea. Everybody in an every executive with an MBA has been taught the beer game or some version of the beer game for decades. And it has taught them fundamentally the wrong thing about supply chain, I think, for the world we live in today. Because basically people take away the wrong thing from the beer game. They take away this idea that like if I'd had better visibility of that signal and I had better predictability of that signal, I would have not had too much beer at the end of the day. Right. And it's just wrong. No matter how people play the beer game, it ends the same way. And the world kind of operates supply chain in that. That's why forecast accuracy is so easy to predict and easy to measure. It's kind of the whole so much of the entire supply chain technical infrastructure has been built around it for 30 years because it's easily measurable. It's the thing every board and a CEO asks about when something goes wrong and you have too much inventory, and you're like, my forecast accuracy was X versus Y, and it's a good bridge. But the problem is that predicting the future has not gotten any better. Like, I mean, you can tune it a little bit, but like you're always gonna be wrong. It's the only thing that's 100% about a forecast is it's wrong.
Ted StankRight.
Dave ClarkJust a matter of how much in which direction. But what you can do is be incredibly more responsive to signal and on information share. And to me, that's the real future of the next decade is the organizations that make that pivot are going to be dramatically further ahead than everybody else. And I would say that's largely how Amazon operated for 20 years was like this constant within the limits of technology, compressing signal. To execution across every single decision.
Ted StankAnd you know the Amazon before Amazon was Dell.
Dave ClarkYeah.
Ted StankRight. And you know, it did a lot of work with Dell back in the 90s, and that was exactly that idea. Now, back then, Tom, getting a computer that we custom ordered online in 10 days, we were we were comfortable with Well, I mean
Dave Clarkit's compression of that whole supply chain and make the order into a really like taking every bit of slop out of that stack, right?
Ted StankSo it's it's a phenomenal idea. I mean it's kind of a simple idea, but I think at the heart of what you're trying to do with Augur, Dave, is what you just said. Compress that signal to response time.
Dave ClarkYeah.
Tom GoldsbyBut Dave, would you characterize what you're undertaking as serious evolution or revolution? Right? Is this something incremental or is this a lockstep?
Ted StankWell, we promised him easy question.
Dave ClarkIt is easy question. I think it's it's for sure a it's a nonlinear progression and improvement because today you
What Future Supply Chain Jobs Need
Dave Clarkmostly don't have a way to unlock that. Like the there's so much value trapped between decisions and between teams. It's where most of the working capital buffers live. You know, it's where massive amounts of like team spend lives, it's where massive amounts of fines and penalties and expedites and everything else live. It is everything that happens when the world disrupts your normal nominal plan. And in supply chain, you know that's like most of the time. And so that is inaccessible largely today without connecting all that stuff.
Ted StankAnd we I I think we've talked about this before too, is when you and I started this b career, we had a a lack, a paucity of information, of data. And what we've done in the 30 years since has been given us tools to create much more data, but now we haven't figured out how to really use it yet.
Dave ClarkYeah, we're like drowning in data and star for insight. Right. You you know everything and nothing at the same time. We were having this conversation a little bit last night. This sort of like go back to that Shaw conversation. There was that whole discussion about supply chain versus military and this sort of dealing with ambiguity and uh what's happening in the field and connecting all people around. Whoever has the high ground in a combat situation wins, you know, historically over time, right? Because they have a better overall connected view of the universe. Supply chain is basically the same kind of thing. It's a it's a big global arena, lots of teams, everybody has their own thing they're doing, disconnected views of reality, individualized goals, and whoever has the best visibility and connectivity has the highest probability of success in winning. And I think what AI is allowing us to do now is to connect the arena from soup to nuts, from commodity to customer in a way that gives the companies that do it the high ground in competing with the And I
Ted Stanklove that metaphor too. Um I just recently, I'm like a military history buff, and I just recently re-watched the 1970 movie Waterloo. And what I failed to realize is that Napoleon was pretty close to winning.
Dave ClarkYou can do other things too. I can't.
Ted StankBut you told me that rowing was lame for me to do
Dave Clarkthat.
Ted StankI don't know.
Dave ClarkNapoleon 70 is not getting better.
Ted StankBut I didn't realize he was about to win until this like massive fog rolled across the battlefield, and he didn't realize that the Prussians were approaching on his flank.
Dave ClarkIt is totally that. It is what you don't see, it's what you don't know, it's what you're unaware of that costs you money, it costs you time, it costs you customers. And that is sort of the game across supply chain every day.
Tom GoldsbyYou know, I I think we harbor romantic notions pursuing perfect data. And and I can tell by your your chuckle, we're not gonna get there, right?
Dave ClarkIt doesn't exist, though.
Tom GoldsbyIt doesn't exist. All right. So how in the reality of imperfect data and all these silos and little bastions, how how do we kind of get good enough in terms of being able to make effective decisions timely?
Dave ClarkWhat I tell people is you you already run these businesses on the data you have. I mean, today, people with spreadsheets and meetings and Power BI and whatever make these decisions day in and day out. Like whatever size, you know, you if you're a $600 billion business or a $600 million business, your businesses run on the data you have. If you and your teams can run on that data, then so can AI. We are in a challenging spot where there's not that many choices for platforming everything with context, but that journey is very doable with whatever data stack you've got. Like a bunch of this is just start. I'm always amazed at how few people are actually doing anything even personally of substance with AI. The number one answer to what AI tools do you use is co-pilot, which basically means you don't use AI. And like if you don't have like bots running across your own personal enterprise, you as a person, then you're just not even in the game. Like you you're just like completely detached from what is coming. And I think what's wild to me is that the distance between what the state of the art is and where the average person is is the widest that I have ever seen. Like it's not even like they it lives in two totally different worlds of what's happening.
Ted StankLet's take this first question, because I think it follows right from this conversation. Uh do you get the impression that people talk about AI in platitudes? Can you share real-life examples of AI agents solving significant problems? I would take that in two pieces in the business world and what you were just talking about in the in the personal world.
Dave ClarkYeah, I think it like we're in the crazy hype cycle, right, like in some ways. So like you have to take some of it with a grain of salt. It's like you can tell who has never used AI by the people who say it will like do everything. And you can see who's never used it or used it once when they say it can do nothing. Both extremes are, I think, wrong. And as with most things, the truth lives somewhere in the middle. I mean, AI does incredible things. I mean, uh in the supply chain world, you know, we have customers that we have taken and moved whole processes into full autonomy where it's not just about uh, you know, when you think autonomy, a lot of times thinks I used to do X, now I do Y. But really, it's how do you take that process and reimagine it, embed optimization science into it, and then execute it. So it's sort of it's not like replacing a task, it's replacing an outcome. Uh and so we see that in a lot of different places. And I think personally, you know, there's now so many choices between like either you know, you can use Muse, you can use uh GPT Dots or Grokbots. Like there's so many ways to set up your own personal team that does work for you every day on your behalf. I mean, I easily do 10x what I could do two years ago. Like I mean, it's not even close. That's the the other fallacy is like I work way more now than I did two years ago. But I get ten times more done probably because w what I find is I get to do more of the interesting stuff because I've got these you got these other things grinding through a bunch of other stuff.
Ted StankYeah. Let's take this one. What what's one leadership story or lesson that stuck with you throughout your career?
Dave ClarkUh probably delegation is the biggest one. I mean, I I remember when I started out, since there's a lot of st students in the room. I mean, when I started out, I was just like I thought I was right a lot. I thought I knew how to do stuff. I trusted myself to do stuff, but didn't necessarily trust other people to do it. And you it took me several years to really get to the place to to really learn and understand and appreciate how much of a order of magnitude lever and capability it is when you truly empower other people to do work, not just to do work, but to understand the objective, to understand the outcome, to be part of that journey in a serious, substantive way. And that you know has done nothing but sort of open the aperture and velocity over time too. And I've just taken that now and sort of added that and coupled now virtual teams into that.
Tom GoldsbyHey, speaking of the people aspect of the business, something that came out yesterday, call out to our friend Alan Amling, who chaired just an excellent panel yesterday. Kind of the takeaway from that
Amazon’s Race For Faster Delivery
Tom Goldsbysession was people process technology kind of in that order. And you know, if we're gonna talk about undertaking a big revolutionary lockstep change in the organization, there's gotta be that readiness. What do you find as you approach a client and kind of assess their readiness and how do you help them to get to a place where they can uh get to the process and technology?
Dave ClarkI think readiness is 100% mental. It's not technical readiness or anything else. It's really just do you believe this is the way the future is gonna unfold? Again, even if you don't know how necessarily, but do you believe this is the way the future is gonna unfold? And do you believe it's important to get started now on that journey? And do you believe you're trying to actually reenvision the way the work gets done, not simply assist people in doing the thing they do today a little easier? If you believe those things, then anybody can do it. But there's a lot of hurdle, you know, change is hard. Uh, you know, it's leaders putting themselves out there. It's risky, right? Because like it's very early days still in the technology side of things. So it takes some special you know, Hutzpah, you know, I think to you know, to do that work.
Ted StankWhat companies do you believe are embracing it in a meaningful way?
Dave ClarkUh I think there's a lot of companies embracing it in in meaningful ways. I mean, you know, we work with you know, Kimberly Clark and Fanatics and uh Onson Minor and a few other folks, and like those are companies that are all, you know, for different reasons in different ways headed down this journey. And Meta, same thing. You know, meta's obviously deep in this space. I think i they come in all shapes and sizes. I'd say the big impetus is usually, as with most change, they have some big problem or North Star, right? Big merger, big turnaround, like they're third place and they want to be second place or first place, or they're sixth place and they want to be third place, like you know, it it's some
Ted Stankburning platform. Burning platform of some kind.
Dave ClarkOr the other one we see a lot is like people have aging technology platforms. So there's a a lot of people who didn't upgrade in the last 10 years to the latest and greatest version of every thing. Uh and so they're kind of at a place of do I do that and spend the next three years doing this piece, or do I do something that skips a step and sort of jumps to skips that reset and goes straight to the next version of what operating system looks like. That's the other place we see a lot of engagement from people. People are trying to avoid they know that they spend another three years on the old that they'll be just three years behind on the new.
Ted StankRight, right. And they're looking at something new anyway, so let's let's jump the curve. Cool. Hey, uh shifting gears a little bit here, we were talking last night that you're gonna get to go to Davos this year.
Dave ClarkGet to.
Ted StankYeah, get to and hobnob with uh the the power players there. A lot going on today in the news about the AI industry, the tech industry. Well, I wasn't sure where you
Dave Clarkwere going with the
Tom GoldsbyI'm still waiting for a question. Well the
Ted Stankquestion is um, how do you see the whole controversy with NVIDIA and Jensen Huang and everything affecting I mean, obviously you're a big client of data centers, I'm sure.
Tom GoldsbyWell, I would throw the midterms into that as well. I
Ted Stankmean,
Dave Clarkhow do we feel about the religion stuff too well? Let's talk about that. You know, look, I I think maybe at the macro level, I think it would be hard for the AI industry to have screwed this up more. Like it's just like the worst messaging humanly possible. They would have been better off to let the AI do the messaging. You know, it's like the only people who I think did it really well was Meta, interestingly. I think you know, Muse, the whole approach to bringing Muse and the Meta models to the public has been a much less, you know, it's been very consumer-oriented. It's very meet people where they're at and help bring them on that journey. Whereas the I think you in the others you have like two fundamental differences, you know, like you have one who's like the I don't know how you mean describe Dario, but like you have this like very like, I am going to be the only company remaining on the planet when it's over, and we will be us and your government, and you know, it's interesting. And then you have ChatGPT, which is like doesn't have a lot of personality, like it's just the other side of it, and it's very kind of feels cold, I think, to the outside. And both sort of telling you that they're going to kill us all. You know, it's a very weird message to say, I'm gonna build this thing. It's got a non-zero chance of destroying all of us, but I can't stop. I have to do it.
Ted StankRight. That
Dave Clarkjust seems like terrible messaging to me. Uh I also think it's just wrong. And so I think they have screwed that up and as a result screwed up the whole data center messaging. And that's why there's this massive revolt against data centers around the U.S. I think that'll resolve itself because there's so much need there, and it will it will be very good for communities. I think the messaging is right to go extract the value from the companies that are bringing the data centers in as part of building the data centers. That seems like a fair and reasonable trade to me. So I think that's where it'll end up and it'll go in the natural progression. There is so much capital in this space that it's dizzying. You know, they throw around numbers that just are bigger than all the numbers that exist in monetary. Sometimes I think they're creating
Tom GoldsbyI mean cajillion. I mean, what
Dave Clarkyeah, I don't know. I think you just have to sort of ignore a bunch of like a bunch of that's just sort of FUD. Like the important thing I think is AI is going to continue to improve. Frankly, for most of what people do today or need AI to do today, the frontier models are already good enough or better than what you need to do lots of the activity today. For most everybody in this room and in most enterprises, what's missing from the models today is not intelligence, it's context and it's platforming to be able to not have generalized intelligence, but have specific intelligence. And so to me, the next few years of real value creation is going to come there. The frontier models are super helpful for like big like drug manufacturer, like you know, the big, big complex problems. But I think the thing that's much scarier, the piece that I think everybody should have in their mind, that is
Why Decision Speed Beats Forecasts
Dave Clarknot like we're gonna be killed by it, but like InfoSec is a totally different game now. Like information security, that whole game is changing. And I don't I we're just scratching the surface on that. And that's why you're seeing banks like you know not allow bots anymore, like you know, because you it's indistinguishable between the user, and so they can't tell whether you took it or they took it, so who's responsible? So, one, you should play with AI more, and two, you should go fully restack your personal virtual security with things like OnePassword and like reset your virtual universe because most everything is like super crackable.
Tom GoldsbyYeah. Hey Dave, could you talk a little bit about I suspect the challenges you've had building a startup organization in pursuit of the talent, competing with the anthropics and metas and all these folks you're out in Seattle? I imagine it's it's gonna be quite a competition to uh enlist the talent you need to build out an organization to.
Dave ClarkA little bit. Because of what we do and because of our history, we have a lot of folks who if you're purely like, I just want to go make a lot of money and say that I work at a really cool place, but that's not kind of our person anyway. The people who come to us actually want to put stuff in production. They have a passion about they get the idea that like the amount of waste that exists in the supply chain is enormous. We have one planet, we have a growing population of people, maybe more importantly than population, we have a growing population of people who have money to spend on things, they are not going to stop spending on things. So we basically have to find a way to supply humanity for the centuries ahead using less stuff. Because people are not going to buy less stuff or consume less stuff. That never works. You can't have a scarcity mindset. So we have to find a way to do more. And that means we have to be better at supply chain. And we have to redeploy people into stuff that matters to be able to do that value creation. And we need to unlock all this financial capital that's currently going into just excess inventory and moving stuff around into stuff that's more value-added in other ways. And so to me, you know, if you're on board with that, you you a lot of people just get that. They understand the power of that. And so they they like the idea of being able to see their work in the world. Uh and that's been very effective for us in finding the the people who are a match of both really, really smart and really, really passionate about changing the way you know the the world actually operates.
Ted Stankhow many people do you have employed? About
Dave Clark145 now.
Ted StankOkay. Yeah, it's growing. Um speaking of that whole scarcity notion, there's a question here about have you seen a company start utilizing ultra fast delivery for reverse logistics? And is there an ROI to it?
Dave ClarkI don't know about ultra fast, but I think you're seeing more and more people use that delivery side for the return. The big thing on those is like consolidation points. Like you know, mostly what I see people trying to do is not get fast, but get consolidated. Like how do they get localized consolidation without having to buy consolidation points, like you know, using retail locations or whatever. Um so I I don't know. I haven't I haven't seen anything specific on that one other than that.
Tom GoldsbyBut meanwhile, I think about your experience at Amazon, right? I mean, your systems were designed for that one-way flow out to the consumer. And yet the reason so many of us start and finish our shopping at Amazon is because we know we can always return it so conveniently, right? And I know Alan and I were out a I think a Woolville distribution center where shoes were coming back in. And it's like, this system was designed to go out, and now we're creating space and processes to accommodate a lot of reverse flows of footwear in that example.
Dave ClarkI think a lot of companies, I mean, most companies, if you sell to consumers in today's world, returns as a and refurbishment is a massive part of your business. I mean, it's especially true in like how you get into things like electronics and apparel and those kind of things. And so I think it is one of the places where I think you're going to see connected systems and AI have an impact. Because there's a lot of allocation, routing, placement, you know, disposition channel management. There's there's a lot of decision stack that's spread across a bunch of different teams in returns. And so I do think that's one of the spaces that is ripe for connectivity and automation and autonomy because just because of how many different teams and functions are. No one terms
Tom Goldsbyto own it. Yeah.
Dave ClarkBecause you need to own it all the way through disposition, because you know uh a lot of times people get good at the first part, but then it sits in a building for months, you know, and like everybody that's just burning capital and and value in the sale on the other side.
Ted StankLet's shift a little bit to people skills and talent management. We've got a question um what skills were must-haves ten years ago that matter less today, and what new skills should people entering the field invest in now?
Dave ClarkI think that's touched on it a little bit with the operations of
Ted Stankthe other.
Dave ClarkI think leadership skills and people management is like not going away. Every minute spent on that is a minute that pays back over time without question. And then I think you just have to embrace where the world is going and start using things and use the tools, use the models. Uh, there's a question here about which one's the best today. It depends. You should use them all. You should use as many of them as you can afford to pay the subscription for, is what I would say. And you know, at a minimum, I would have Claude, GPT, and grok running. Like grok or Muse are the best for your personal agent stack. You know, if you're a heavy meta product group, especially, you know, WeChat and Instagram and all those pieces, like Muse is can be a really fantastic product. Um and then Grok is great on the other side. Grok tends to be a little more like software developer side with cursor, and Muse tends to be a little more consumer side. But if you don't have those things running, you should do it in between sessions today.
Tom GoldsbyYeah, you know, I have to say, yeah, I use it daily for for various purposes, and I'm usually using it for things
Imperfect Data And Getting Started
Tom Goldsbyeither I just don't want to do or don't have time to do, or just need a little added insight. But the hallucination factor is real, right? I mean, the number of times I go back to my Gen AI tools and I use multiple ones, like, listen, you got this wrong. And it it's funny to get that apology back, you know, but it routinely comes like, no, I wanted the aisle seat, you know, and you put me in seat number 16 claiming it was an aisle seat. Okay, your apology doesn't mean much to me today, but I've informed the next user that, you know, seat number two is probably the aisle seat. But you know, how how do you kind of help us all to kind of gain confidence? I mean, uh how much use in trial and you mentioned the the the criticality of of context is so important, right?
Dave ClarkOne, I just wouldn't worry about it. I just keep doing it. Just keep going. I mean, it's early. I one, I would not expect it to be perfect right now. In fact, it's some of it's bad. A lot of it's really good. If you went four months ago, a whole lot of it was really bad. Like four months ago, you you didn't have a gentic deployment in this way on a bunch of stuff. And so a lot of this is just understanding and being current with the art of the possible, right? And so the hallucinations today are so much less than they were four months ago. It's wild. It's
Ted Stankcrazy that you're talking in four months into four months.
Tom GoldsbyFour
Dave Clarkyears. I
Ted Stankmean,
Dave Clarkwe s we basically the way we talk about it is like, you know, every day's a week, every week's a month, and every month's a year. You know, is kind of the pace of movement.
Ted StankAnd, you know, this is kind of analogous to the conversation about autonomous cars, right? When an autonomous car has an accident or kills somebody, the media and the public just go crazy. But how many times does a drunk driver, you know, kill somebody?
Dave ClarkOr just somebody. Or just
Ted StankI
Dave Clarkmean it's the same thing as supply chain. So how
Ted Stankmany times does a travel agent put you in the wrong seat?
Dave ClarkI've had assistants put me in the wrong seat plenty of times over the years. But like you also have this there's a bit of the We compare agentic outcomes to the perfect person.
Ted StankRight.
Dave ClarkRight. Like human teams do a lot of stuff wrong.
Tom GoldsbyFair enough.
Dave ClarkAt least has been my experience. And you know, the wide distribution of outcomes across a group of people. Whereas the machine, you get a narrow distribution, it's either really wrong or really right. And you know, you you can move the mean, and once you move it, it stays. So I think there's also a bit of you've talked about romantic notions before. There's a little bit of romantic notions around the idea that somehow or another humans are more auditable and more predictable than the machine can be. And I just think that's people can give you an answer. That does not mean it's I've been given many answers that were not the real answer. Hallucinations from people. Yeah.
Tom GoldsbyWell, I think there's a human tendency. We want to be able to lay blame on someone. Well, you want this. You want to announce somebody named Grok.
Dave ClarkSomeone's always accountable.
Tom GoldsbyRight.
Dave ClarkThe accountability is important. I think that it's less satisfying. I I would say it's less satisfying for weak leadership. Because if your leadership depends on blame, then you're you're kind of not that good at it. And so if you need somebody to blame as opposed to looking at the process for improvement and being clear about the the whys, then AI is going to be very unsatisfying because it's harder to fire. Whereas, you know, people who you know you can be a little lazy in leadership and be like, oh, I just, you know, my planning was bad, I'm gonna get a new planner.
Tom GoldsbyWell, let's talk about that just a little bit. I mean, what leadership skills do we need to adopt in the face of the technology and its capabilities in terms of what we're gonna be expecting of our team members moving forward? It seems as though leadership is gonna take on a different orientation too.
Dave ClarkYeah, I
Tom Goldsbythink.
Dave ClarkLeanne says this a lot. People are are more important in a world of AI, not less. Because you really have to help people it's not transitioning to a world of like, in my view, like a world of like lesser. It's like but viewing the world in outcomes across a process and how I manage towards that, and how I manage a system is a very different process management than if my job is to make sure this shipment gets through these processes to this outcome, right? It it's a much more thoughtful, problem-solving, richer environment. And you know, I think uh if you've led people, you know all change is bad. Even good change is bad. And because it's new and new stuff is scary and new stuff is always risky, and so peop you you have to like guide teams through it. And so the leaders who do this best and the leaders who will get the most gains and will get the most self-gains over the next decade will be the people who are the best at you know bringing their teams along that journey to get the organizations to move on that journey.
Ted StankI think there's another nugget in what you said, Dave, and
Security Risks And Real AI Adoption
Ted Stankit's complimentary to a panel we had yesterday about um customers and focusing on the customer, is that we need to focus more on the ultimate outcome. You know, again, we are so siloed, so everybody's got their little metrics. And you gotta pick your head up and see that total process.
Dave ClarkIf you ask the CEO what their measurement is, like revenue growth, margin improvement, working capital reductions, goods to cash reduction. Like they're not saying we need our forecast rate to be 10% better. And so what you want to do is move the entire organization's output to be focused on delivering that outcome in total across the system.
Ted StankBefore we break, we got a little bit less than five minutes. I I think there's a good one here. We've talked a lot about using AI, and I'm sure there's people in the room that are like, you know, I I know I need to, but how do I get started? Are there good resources for people to go to that kind of give them a how-to on how to get started with, particularly with agentic AI and bots?
Dave ClarkI would ask the bot. I mean, I'm serious. Like what I what I would do, like for yourself, what I would do is I would go into Grocker Muse or one of those, and I would say, I want to build a personal chief of staff and team to help me with X, Y, and Z and to run this. What should I do? And you will get a whole plan. And in fact, most of them will actually then say, Would you like me to execute this plan for you? And then they will create the bots for you and frame them up exactly how you want. You know, my advice is don't give it your password to your social media account or your
Ted Stankbody.
Dave ClarkEverything else is good. But I don't have you can't have my Twitter account or my bank account. Those are both too dangerous. And you know the bank won't give you your money back and people won't give you your reputation.
Ted StankThat's a perfect example of how we need to shift our thinking, right? It's like that example of me having to go to the UPS store to mail the computer charger instead of saying, wait a second, let me get on my phone and order it from Amazon in 30 seconds and have it delivered. If
Dave ClarkI want to learn, if there's something that I don't know and I want to learn, I I'll get on and say, teach me and build a curriculum and pull up all the most ranked videos and content and what have you, and you just do it in an hour and be done.
Ted StankAll right, I'm gonna expose us, Tom.
Tom GoldsbyOh my.
Ted StankDave, you've been a partner of ours for a long time. You're a you're a graduate of our programs. You keep coming back, you've hired a lot of our students. What is it we do well and what should we do better?
Dave ClarkLook, I think this forum is a great example of what the university does really well. It's like the sessions yesterday, right? Like they're not like ethereal, high-minded, like academic in a closet conversations. They're like, how do you actually do things in the real world with real tools? And that's what I took from my education here. I think it's what I continue to see the graduates that come from UT having. Every time I come back, I see them getting that experience. And I think that's special. Not enough universities really prepare people for the world they're going to walk into and to be successful in the jobs they're likely to walk into. And I think that's a place UT does really well. If I were to say do more of something, if I look at all the content that's produced externally that I see from UT, the amount of that content that lives in this decade in terms of technology is small. And I think there's a tremendous opportunity to lean harder into that space and into like what's coming
Building Agents Into Education
Dave Clarkand what's there in a way that uh could position the university into a big to be fair, like I it's a most people are in that book. So you're not you're you know, it's not like you're sort of thumb, but it's a it's a it's a way to get again to there should be a program. You should not be able to exit the school without having agents running as a student. When you set up a company in a case study, you should be setting up an agent team to run that company. Like you know, like it's how do you integrate those things into curriculum? How do you create opportunities for people to really learn how the tools actually function in an environment like that? Because the gap between students who come out with that and the people who are in industry and aren't messing with it at all is going to be enormous. And it's a place where the best and brightest and most aggressive will be able to do dramatic pivots in velocity in their career.
Ted StankI think that's something for all of us to take to the bank. Absolutely. Great stuff. Dave, as always, it's uh it's always wonderful to be on stage with you. Your brain just works in ways that amaze me all the time.
Dave ClarkThanks for having me.
Ted StankEverybody, let's have a big hand for him. He's uh, I don't know, are you staying around today?
Dave ClarkFor a little bit.
Ted StankYeah. Okay.
Tom GoldsbyAll right.
Ted StankThanks, man.
Tom GoldsbyGreat stuff, man.
Intro & OutroThanks for listening to Tennessee on Supply Chain Management. Be sure to subscribe on your favorite podcast platform for more conversations with the leaders framing the path for practitioners, scholars, and students. Have a question or idea? We'd love to hear from you at gsti at utk.edu. Until next time, listeners.