Harnessing AI for Smarter, Faster Supply Chains with Steve Hochman

[00:00:00] So one of the things we're seeing in AI is just you can get a huge amount of benefit by just running very small scale experiments.

[00:00:07] The key is creating the conditions and that's what companies are doing.

[00:00:10] And the idea is that they want to foster experiments in an organized way so that the successful ones can scale quickly.

[00:00:16] Those types of companies are getting a lot of value by figuring out how to do experimentation well with AI.

[00:00:21] Hello and welcome to the Next Level Supply Chain with GS1 US, a podcast in which we explore the mind-bending world of global supply chains,

[00:00:29] covering topics such as automation, innovation, unique identity, and more.

[00:00:34] I'm your co-host, Reid.

[00:00:36] And I'm Liz.

[00:00:37] And welcome to the show.

[00:00:39] So today, Reid and I had a conversation with Steve Hochman from Zero 100.

[00:00:46] And while many of the folks that we work with touch the supply chain,

[00:00:51] Steve is really in the supply chain and shared with us what is happening today

[00:00:58] and what will help get organizations to the next step in their supply chain efficiencies.

[00:01:05] It was a really cool conversation where sprinkling in artificial intelligence can really be helpful,

[00:01:12] but focusing on the key components that we've been talking about for years, people, process, and tech.

[00:01:19] It was a great conversation.

[00:01:20] Steve, welcome to the show.

[00:01:22] We are so excited to talk to you and learn a lot more about AI and all of the fun kind of emerging technologies that we keep hearing about.

[00:01:32] Nice to see you and great to be here.

[00:01:34] Excellent.

[00:01:35] So before we kick off and really get into the nitty-gritty,

[00:01:38] tell us a little about yourself and your background and about Zero 100 and just set the stage for this conversation.

[00:01:45] So first off, Zero 100 is a research organization that serves specifically chief supply chain officers

[00:01:50] and chief operating officers at large global companies trying to answer one simple question,

[00:01:56] easy to say, hard to actually answer, and that is what is the future of supply chain?

[00:02:01] And part B of that question is how do I capitalize on that future for my organization?

[00:02:07] We live in data, so we're always reading trends to understand what that future might look like,

[00:02:13] recognizing that there's no playbook for plans that are in the future.

[00:02:16] And so that's the community that we serve.

[00:02:19] My job as VP of research is to help set that research agenda overall for the organization.

[00:02:24] I have some specific functional areas of specialization, things like supply chain planning

[00:02:28] and other functional topics like how do you actually upscale people, really tactical stuff.

[00:02:34] But big picture, what we're trying to do is pull together all of the major capability themes,

[00:02:40] back those themes with data to be able to help supply chain leaders figure out how to

[00:02:44] build towards a winning future.

[00:02:46] My background, 20 plus years as an operator, as well as consultant and analyst.

[00:02:51] 10 of those years might be the most recognizable brand name would be Nike,

[00:02:54] where I led supply chain strategy for the company and also played a couple of leading P&L roles in supply chain,

[00:03:00] but also a startup entrepreneur. I like to think of myself as a supply chain entrepreneur.

[00:03:06] And typically those words don't go together, but that's very much how I think of my personal brand.

[00:03:10] So this is fun. It's great to be able to have a conversation about the art of the possible.

[00:03:15] It's what we do for a living, but also I gather the topic at hand today.

[00:03:18] And as Liz said, we are excited to dive into this.

[00:03:21] And I know AI is going to come up and we will talk about that.

[00:03:26] But we're interested to know, are there significant trends that you're seeing in the industry right now?

[00:03:33] And or are there emerging technologies or innovative ways that are causing disruption?

[00:03:41] It doesn't feel like 2020 was that long ago, right?

[00:03:45] And, you know, January 2020, everything's thumping along, going really well.

[00:03:50] And then March, boom, we hit COVID and the world changed.

[00:03:54] But be interested in your perspective here at 2024.

[00:03:56] Yeah. I mean, I think some of this hopefully is not a surprise to those who are reading the newspaper in terms of the why.

[00:04:02] But just to reiterate before talking about a specific trend, it's what's going on behind the scenes is volatility, right?

[00:04:10] And change. So pandemic was a was a spike.

[00:04:12] But what we're learning is that the spikes are coming from all angles, whether that's geopolitics or public health or environment.

[00:04:19] We used to talk about black swan events as this sort of special newsworthy moments.

[00:04:24] And now it feels like you turn the Wall Street Journal front page over and it's black swan number two or number 12.

[00:04:31] Right. So that's that's a little bit of the environment that supply chain leaders are living in.

[00:04:35] And the result is that people are asking the question, trying to build towards agility and enterprise resilience.

[00:04:42] Those are kind of keywords that you'll often see show up in strategy documents from some of the companies I just mentioned.

[00:04:48] And what that means from a trend perspective is and I like to use the word capability because it's definitely technology is a component of it.

[00:04:55] But thinking about the people processing the tech together, the companies are trying to build a muscle, if you will, to drive that kind of agility and resilience, which is hard enough when you're, say, a startup, five million dollar company.

[00:05:09] But when you're Walmart and operating a complex global network or if you're, you know, just on the retail side where you're Honeywell in diversified industrial products, the complexity of different product lines, channels, geographies.

[00:05:25] Verticals.

[00:05:26] Yeah.

[00:05:27] Yeah.

[00:05:27] It's the proverbial problem of how do you help the elephant dance, right?

[00:05:31] Like it's, you've got all that complexity, which is part of the differentiator combined with the volatility and you have the recipe for challenge, but also therefore innovation for those who can actually crack a code.

[00:05:43] But at a macro level, it's about that agility and resilience.

[00:05:46] Now, if you break that down and you say, okay, well, what are the building blocks of that that are emerging?

[00:05:51] Some of it was spurred by the pandemic for sure.

[00:05:54] When you think about the sort of need for hyper agility in that moment, but still continuing because of all those other forces at play, there's a drive through the lens of process first.

[00:06:03] And I'll talk about tech last.

[00:06:04] So the idea that you could be historically, maybe pre 2020, you'd be really good at demand planning or really good at logistics and get rewarded for that huge lift in the business, huge lift for your career.

[00:06:16] If you want to really be as agile as we were just talking about, you can't actually afford to just be good at a function.

[00:06:22] You have to be good at cross-functional or systems level change, but also systems level adaptation on the fly.

[00:06:27] And so this notion of connectivity is something that was talked about, but it's very much connectivity on steroids.

[00:06:33] But the other is not just the ability to connect the dots between functions, but it's also the cycle time.

[00:06:40] If you're feeling whatever that disruption is, and by the way, I think today we were reading about the potential for a port strike on the East Coast, right?

[00:06:47] So pick the moment.

[00:06:48] The minute those next disruptions are on the horizon as they inevitably will be, the ability to not just connect people, but iterate on a plan in real time is just as important as the connection between those individuals.

[00:07:02] So it's the pace, the cycle time that's also changing.

[00:07:05] So speed and scope all at once, those are all fundamental shifts that are happening.

[00:07:10] The third, of course, and this is leading the witness a little bit around artificial intelligence, but is the intelligence attached to A and B, right?

[00:07:17] Like the ability to power those more complex systems with information that used to be done on spreadsheets with maybe, if you were lucky, you had some data scientists in-house who could do some homegrown algorithms, but in a local pocket of the organization.

[00:07:32] But now algorithms that are connected and factoring in, thanks to artificial intelligence, a much broader range of variables.

[00:07:40] So not just if we're thinking about supply chains, say, for example, the history of orders, but the weather forecast or the latest social media posts, range of data inputs into those algorithms.

[00:07:50] So both connecting algorithms, but also enriching them with far more data sources for the purposes of intelligence, right?

[00:07:58] But intelligence for purpose, right?

[00:08:00] In that case, it's really supply chain.

[00:08:01] So those big three, the scope, the speed and the intelligence are arguably the big three shifts that we are continuing to accelerate.

[00:08:10] And that's what we see happening.

[00:08:12] When you talk about those, because those, it sounds so simple, and I know it's not.

[00:08:17] When you talk about putting teams and processes together, are you talking about internal teams and processes?

[00:08:24] Or are you talking about a connected supply chain where you have teams from various organizations talking together?

[00:08:32] And especially, you know, you talked about the intelligence and that connected systems.

[00:08:37] Is it a supply chain talking together and internal, like the whole thing?

[00:08:44] So there's multiple levels of collaboration that are occurring, right?

[00:08:47] So there's already been in that simple statement that I made.

[00:08:50] One is certainly within functions, right?

[00:08:52] So the ability to align on, and I'm going to use planning as the sort of the poster child for this, but align on the inputs to a forecast.

[00:09:00] That's a collaborative exercise that happens within a function, and that has to occur more intensively in this situation we're describing.

[00:09:08] And if you think about the expanding circles, it's also balancing supply and demand.

[00:09:13] So that means collaboration across parties within the supply functions, demand functions.

[00:09:17] That used to be called sales and operations planning, and that was amazing, and it yielded tens of billions of dollars in shareholder value.

[00:09:24] But now that's not enough because, again, of all of those other factors we just discussed, not to mention competition is doing this.

[00:09:31] And therefore, you've got to think about how do you involve the commercial organization in that same plan, right?

[00:09:36] And then add to that, if you're, say, Procter & Gamble serving Walmart, how do you actually assure store shelf availability?

[00:09:44] It means you've got to collaborate planner to planner or either planner to logistics person or buyer, right, across price boundaries, and the same thing on the supply side.

[00:09:52] So it's an and, and, and.

[00:09:53] And to some extent, it's, to your point, it's not easy, right?

[00:09:57] Because the more you do that, absent the tool sets and the processes and the mindsets, the more risk there is that you're just going to grind yourself to a halt, right, with all that complexity.

[00:10:07] But, in fact, that's the basis of competition.

[00:10:10] To some extent, yes, there are opportunities to simplify.

[00:10:13] That's always a good thing.

[00:10:22] Ecosystem that we're now managing.

[00:10:23] And that's the basis for competition.

[00:10:26] Complexity management.

[00:10:27] It does.

[00:10:28] It does.

[00:10:28] I'm actually thinking through, but it's, we're going to dive into this more because the scope, speed, and intelligence, you know, it all kind of comes together in that.

[00:10:37] But, like, I'm just thinking of, like, planners, logistics, buyers, and communication streams, visibility, over communicating, under communicating.

[00:10:47] There's a lot to peel back.

[00:10:48] And we're doing our own business planning.

[00:10:51] And, you know, our different teams are moving to agile, you know, workflows, which are great.

[00:10:58] But it's also, like you said, you're developing a new muscle.

[00:11:02] 100%.

[00:11:02] You develop that new muscle.

[00:11:04] It is very sore and painful the first few times.

[00:11:07] No pressure.

[00:11:08] Painful.

[00:11:09] And then you kind of get into it.

[00:11:10] We talk about, like, unconscious competence, right?

[00:11:15] Like, as human beings, I stand up.

[00:11:18] I don't think about balancing or putting extra pressure on my heel or my toe.

[00:11:21] I just do it.

[00:11:23] Breathing.

[00:11:23] And that's what we're trying to get to with businesses as well, with macro level things, like you said, Liz, like real high level, like collaborate, make sure that we're talking.

[00:11:33] But it's also, how are you talking?

[00:11:35] How often are you talking?

[00:11:37] Is there a purpose to your talking?

[00:11:40] Are there action items that you're walking away with?

[00:11:43] Or are you just doing readouts?

[00:11:45] And sometimes readouts sound like the old 1970s, 1980s Charlie Brown when the adult came in the room and you heard wah, wah, wah, wah, wah, wah, wah, wah, wah.

[00:11:56] I guess that resonates.

[00:11:58] At the end of the day, and it's good.

[00:12:00] It's partly why I haven't talked about technology yet.

[00:12:02] It's got to be human-centered because we are who we are.

[00:12:05] We're evolved to be this species that operates this way.

[00:12:08] And when you start to layer on the complexity of lots of humans, even if you're all physically in the room, it's hard enough.

[00:12:13] Now you're talking about time zones and everything else.

[00:12:15] It seems overwhelming.

[00:12:17] But I guess I'm here to say, right, that companies are managing through that and getting better at it despite all of those challenges.

[00:12:23] That's what I think is interesting is that, yes, we put the problem on the table, but we're seeing movement.

[00:12:29] The question is, like, are there trends?

[00:12:30] The answer is yes.

[00:12:31] There's trends in actually tackling this very human problem with supportive technology while also supportive leadership.

[00:12:38] And so there's absolutely both conversations.

[00:12:40] So you did talk in the very beginning, people, process, and tech.

[00:12:45] They all need to come together.

[00:12:46] And I've heard this multiple times.

[00:12:48] You can have the best technology, but if you don't have support of the executive staff, because even the best laid plans run into reality.

[00:12:57] And reality needs to be twisted, changed, tweaked, and responded to, reacted to, because it's never simple.

[00:13:06] It's never easy.

[00:13:08] There's always something else.

[00:13:09] And you need the people support, emotional, financial, time, willing to carve out other things to develop these new muscles and then to go about it.

[00:13:19] Here at GS1US, we've had our IT staff come on our all-hands meetings in the last couple of months talking about co-pilot.

[00:13:27] We leverage co-pilot.

[00:13:28] And people are sharing what they're doing and how they're leveraging and how they're using it.

[00:13:33] And I'm like, that's the new way.

[00:13:35] It's just, you know, for us older folks, it's kind of like when we went from pagers to cell phones and then email to online orders to getting rid of fax machines.

[00:13:46] And it's just, it's this evolution, but you have to keep up and you have to try new things and you have to experiment.

[00:13:53] It doesn't mean that the latest and greatest is the best for your organization.

[00:13:56] But if you stay there, you turn into a historical site.

[00:14:01] You know, like we still don't heat by fire in our homes too much.

[00:14:05] And some of us like fire, but you know what I'm getting at, Liz.

[00:14:08] Yeah, it's great.

[00:14:09] And I think there's a little bit of a history part of this, right?

[00:14:12] Where disruption has happened in industry before because somebody was moving too slowly, right?

[00:14:16] That's not new.

[00:14:17] Like we think about, you know, way back in the days of, say, circa 1994, when the internet was kind of a twinkle in our eyes.

[00:14:24] We had this idea that there was this thing called the World Wide Web.

[00:14:27] And Jeff Bezos was sitting in his, I don't know if he was in his basement or top floor of his house, but, you know, in his startup, thinking about not is the World Wide Web a marketing vehicle, which is what most of the world was thinking of it as.

[00:14:40] Hey, how do I put my store billboard up on this thing called the World Wide Web?

[00:14:43] But how do I use it to change the customer experience?

[00:14:46] That different mentality was because he was that guy going and kicking the tires on the new tools and saying, what if, right?

[00:14:53] Yeah, and he was working for a financial institution and he did the math and he's just like, if I got 2% of this, I'd be a billionaire.

[00:15:00] Right, exactly.

[00:15:02] 2%.

[00:15:02] Not everybody can be Jeff Bezos, but the idea of the curiosity and the willingness to experiment and to put in the extra maybe a couple of minutes even in your day, that behavioral norm is at the center of what we're talking about here.

[00:15:15] When it comes to, I'm agreeing with your point, right, which is be ready for whatever's coming because we don't actually know.

[00:15:22] But if you can be curious, you're going to be better off.

[00:15:25] And we actually have statistics on that.

[00:15:27] It's really cool, like the forward thinking and read to your point, like just not getting complacent too.

[00:15:32] And for each use case, there's going to be a different solve.

[00:15:35] It might not be the most leading edge tech.

[00:15:37] It might be somewhere else, but knowing what else is out there.

[00:15:40] Steve, I wanted to just ask you, if we think about AI, because it's been on the tip of our tongues probably this whole however long we've been chatting, 20 minutes,

[00:15:49] what are some of the most exciting things, developments, items that you've been seeing from an AI perspective going on in industry right now?

[00:15:58] When we answer this question in general for our customers, I'm sort of, I'm playing that back to you a little bit.

[00:16:03] The research we do points us to some who are yield, getting benefit now from a lot of quick wins in focused functional areas.

[00:16:11] We'd be remiss if we overlook that.

[00:16:14] I want to come back to that.

[00:16:15] There's a ton of value in vast array of small experiments that are going on right now in every organization that is thinking about this for exactly the reason that Reid was just describing,

[00:16:24] where people are trying to learn the technology, but also to create momentum behind bigger transformations, right?

[00:16:30] So one of the things we're seeing in AI is just you're not deploying some huge ERP system.

[00:16:35] You can get a huge amount of benefit by just running very small scale experiments.

[00:16:40] The key is creating the conditions, and that's what companies are doing, the good ones.

[00:16:44] I mentioned P&G before, so just to pick up on that example, they have an AI factory.

[00:16:49] They actually call it that.

[00:16:50] And the idea is that they want to foster experiments in an organized way so that the successful ones can scale quickly, right?

[00:16:57] So you're not just running a bunch of science experiments.

[00:16:59] The commitment to creating that ecosystem for experimentation in AI is a big idea.

[00:17:04] Combined with the instrumentation of the process to say, you know what, we're going to make sure that this doesn't just end as an experiment, but we're going to scale up the successful.

[00:17:14] That one-two step in AI is really translated to huge value and also competitive separation for companies like P&G.

[00:17:23] I mentioned Honeywell, et cetera.

[00:17:24] Amazon, of course, is the poster child for this, right?

[00:17:26] Those types of companies are getting a lot of value by figuring out how to do experimentation well with AI, number one.

[00:17:31] Now, the other way we answer that and what we track is looking at those disruptors and what are they actually doing with AI that's making a difference to shareholder value, to customer delight, to sustainability, the metrics that matter for enterprises.

[00:17:47] And what we see there is really on the other far end of the spectrum is companies who, what we might call AI natives, sometimes out of Asia, actually.

[00:17:57] Some of the e-commerce ventures like the Alibabas of the world, Shein and Fast Fashion, who are, despite the important asterisk around what impacts they're having on the environment, et cetera, with Fast Fashion, which we can come back to.

[00:18:09] From an AI perspective, what we see is companies, and it happens to be concentrated in Asia because of some other cultural norms and some precedent there, that are really taking a blank slate view of business processes and saying, how could AI help us serve customers better?

[00:18:27] Just like Jeff Bezos was doing with the World Wide Web way back then, they're asking that same question about business models writ large with respect to AI.

[00:18:36] And they're essentially, I'll say, semi-automating processes that were not thought automatable and driving not just local optimization, but global connected optimization by creating essentially a mesh, right?

[00:18:50] A data mesh.

[00:18:50] And then using algorithms that are connected to run processes from everything from, in Shein's case, from personalized offers to individual online customers, all the way back through to capacity planning decisions, back in an ecosystem of suppliers, 6,000 suppliers in Guangzhou, China, and connecting them all in real time and optimizing decisions around, for example, what should I sell that customer online right now amongst 35 million customers?

[00:19:19] Based on the capacity I have available up here.

[00:19:50] Again, dynamically all in real time.

[00:19:51] So the reason I picked the Shein example, and we talk about this, we wrote a report on this recently, is if you look at the metrics that are attached to this from a business perspective, it's really quite stunning.

[00:20:01] So just to throw this out here as an illustration of the from and to, if you think about a traditional fashion company that doesn't run on AI, that's very successful.

[00:20:11] You know, they may introduce 150 styles if they're good in any given month.

[00:20:17] Sometimes they'll do faster, like, you know, for certain seasonal launches, it could be 150 a day even.

[00:20:22] But it's that order of magnitude.

[00:20:24] If you look at what Shein launches per day of new styles, 6,000 per day of new styles, brand new styles.

[00:20:32] And they're also able to offer it at extraordinary low price.

[00:20:35] Because, again, they're running these processes, not just with intelligence, but also with extreme efficiency.

[00:20:42] So very low headcount for very high agility.

[00:20:45] Wow. I'm completely floored by that.

[00:20:48] I want to keep us moving because we're going to keep peeling back the onion here.

[00:20:51] But I want to talk about the data itself, because we've all heard that data is the new oil and that's been around for 20 years now.

[00:20:59] And companies, a digital transformation and digitization.

[00:21:03] And, you know, even in our own businesses, you know, it's like, well, I'm feeling at this.

[00:21:07] OK, well, I need more than feelings.

[00:21:08] I need facts.

[00:21:09] I need data.

[00:21:11] How important is data?

[00:21:12] Do you have any concerns and are we seeing the right trends like that are happening?

[00:21:19] Are you seeing, you know, some bad trends around data where it's, you know, because garbage in is garbage out.

[00:21:24] I always remember that one.

[00:21:26] So the dependency is real and intense on data for this AI discussion in particular, but also more generally for digital capability.

[00:21:34] Right. So but in AI shines a light on the vulnerability.

[00:21:36] If you don't have good data management processes and capabilities more broadly, then it is exactly what you just said.

[00:21:43] Garbage in, garbage out.

[00:21:44] And so companies, they were able to skirt by, you know, so skate by with some kind of bootstrapped approach to data governance.

[00:21:52] When they start to layer on some of those sophisticated algorithms we just talked about.

[00:21:57] Scale aspects.

[00:21:58] Yeah, it just it becomes a nightmare very quickly.

[00:22:01] And that's true for even this.

[00:22:03] So this is a pressure point, even for those companies that would consider themselves advanced.

[00:22:07] So we have in our networks and companies who themselves are high technology companies who are known for being excellent at data management.

[00:22:15] And I heard from one person who was the chief supply chain officer who described this as a process from a data management perspective,

[00:22:22] as constant gardening, which doesn't sound very high tech, right?

[00:22:26] No, but it makes sense.

[00:22:28] It's a huge problem and opportunity for those who can figure out how to streamline those processes and make data accessible.

[00:22:36] And so it's a rich topic for sure.

[00:22:38] If you're interested, we talk about sort of the breakthroughs there, but there are breakthroughs.

[00:22:42] It's just it's we're very much in the middle of that.

[00:22:44] I love that analogy, constant gardening, because that resonates and that is the reality.

[00:22:51] Like you have to be tending to it all the time.

[00:22:54] This is not a Ron Popeil set it and forget it type of thing.

[00:22:58] This is weeding, watering, fertilizing, soil pH samples, all of that when you're really getting into it.

[00:23:06] It's not sexy.

[00:23:07] It's not fun.

[00:23:07] It's not money generating.

[00:23:09] But at the same time, without it, it's going to hinder.

[00:23:12] And watching conversations, I have a lot of data quality conversations in my life because I'm really passionate about it.

[00:23:18] But it's so important to have that good foundation in all of the data that you have.

[00:23:23] And I do.

[00:23:24] I totally agree with you, Reed, because constant gardening to have a beautiful flower.

[00:23:28] It takes all that time and attention and you're bending over and your back hurts and all the things.

[00:23:34] But you have to.

[00:23:36] You got to talk nicely to it as well.

[00:23:38] Yes.

[00:23:38] Yeah.

[00:23:40] Yeah.

[00:23:40] No, I think it's awesome.

[00:23:41] Okay.

[00:23:42] See, we've talked about data.

[00:23:44] If a business wants to start thinking about AI and, you know, you gave us that example of P&G, which I love because it takes, you know, that paradigm shift.

[00:23:54] Try little things.

[00:23:55] Not all of them are going to work.

[00:23:56] That's totally cool.

[00:23:58] How would you suggest a business would get started with AI?

[00:24:02] Is it good data?

[00:24:03] Is it test and learn?

[00:24:05] Like if they're starting from that blank slate, like you mentioned, that they've been doing.

[00:24:09] So what we see happening amongst companies who are cutting their teeth on this is absolutely picking up on this conversation, right?

[00:24:16] Making sure that if you use the language of minimum viable product, that you have minimum viable data for particular experiments.

[00:24:23] But it's also being super clear and surgical around what problem you're solving.

[00:24:29] There's a sort of a left and right or yin and yang, pick your analogy, balance between on the one hand wanting this imperative that we get our data right and on the other being super focused about where and when we actually put our efforts into data cleanup so that we're not boiling the ocean.

[00:24:44] Finding a few experiments that you understand are going to add value if you address the problem and then getting the data right for those.

[00:24:51] So that's a piece A and B, focus and get the data.

[00:24:55] Even if that process of acquiring data is scrappy, you can manage it on a small scale usually.

[00:25:00] What that yields is number three, which is the intangible, which is momentum.

[00:25:04] When you actually add value because you're solving somebody's real pain point, then you get permission to go continue, right?

[00:25:11] And this is the real genius at Agile.

[00:25:13] It isn't so much that you're reducing cost of implementation, although it's part of it.

[00:25:18] It's that you unlock the psychology of belief, right?

[00:25:21] Because people are like, oh, I see it, right?

[00:25:23] You build something and you add value.

[00:25:25] So those three are true for AI.

[00:25:27] And again, I think the benefit of AI is there doesn't have to be some huge systems implementation.

[00:25:31] That's generally the flywheel that we see companies implementing to be able to create larger and larger range of experiments that are faster click.

[00:25:40] But I'd like to get your perspective on the ethical implications of AI.

[00:25:46] And this one is unique to me, and it kind of takes us a little off to the side.

[00:25:49] There's a few things I just want to kind of get out here in my head right now on AI.

[00:25:53] In 1994, like you mentioned, World Wide Web kind of came about.

[00:25:58] It was already there, but that's when it kind of hit mainstream, and we were seeing it on the Today Show, if you will.

[00:26:03] In 1997, I started working for Cisco Systems, the internet company, and we were building websites and data centers left, right, and center.

[00:26:13] And it's still, we didn't have video then.

[00:26:16] You know, we didn't have telephony on it then.

[00:26:19] We didn't even have point of sale sales transactions happening then.

[00:26:24] Not in the big case that we are.

[00:26:25] So that was four years.

[00:26:27] Then it was really like 8, 10, 20, and, you know, we're still here and we're still developing.

[00:26:32] But I think of AI as like 2022 with ChatGPT, right?

[00:26:37] I think it was either 2021 or 2022, November, when it kind of launched, and we're already on like version 4 or 5 now.

[00:26:44] But the ethical piece comes to this.

[00:26:47] I remember being in the office one day and talking about it, and I was like, oh, I played around with it.

[00:26:51] And I actually asked it to write a letter to my wife, and I gave it some background.

[00:26:55] And the person said, like, look, don't you feel like you're cheating or it's unethical?

[00:27:00] And I said to that person, I was like, that never crossed my mind.

[00:27:05] But don't you go to the pharmacy and buy, you know, Valentine's Day cards that somebody else wrote, and then you write your own little thing in?

[00:27:14] I said, so I kind of see this the same way.

[00:27:16] I gave it prompts.

[00:27:17] I put things in, and then I changed things.

[00:27:21] But it helped me get 80% of the way there in 10 seconds versus me flumbling for hours.

[00:27:28] It made you better at what you do.

[00:27:30] Right.

[00:27:31] Versus trying to replace what somebody else does.

[00:27:34] And I think that's where the misconception is.

[00:27:36] Yeah, 100%.

[00:27:37] So there's, yes, there's many dimensions to the ethical question, right?

[00:27:40] But in terms of if part of the ethical question is, are we cheating, if you will, by doing things more simply?

[00:27:47] Yeah, no, I actually, I'm with you.

[00:27:49] We need to be transparent about what we're doing.

[00:27:52] A lot of black box AI out there.

[00:27:54] Yeah.

[00:27:54] We should not confuse transparency with ethics.

[00:27:57] Like, if you have, or I should say innovation with ethics, if we're actually innovating because we're actually delivering a better Valentine's Day card, right, and somebody gains more happiness from it, then congratulations.

[00:28:10] Right?

[00:28:11] Or you're able to deliver three Valentine's Day cards to your wife instead of one.

[00:28:15] But, yeah, that's progress in my view, right?

[00:28:18] Now, and I think that is probably shared with a number of people.

[00:28:21] The trickier part is when you're not transparent about the approach that you're taking, or even worse, we're using lack of transparency to manipulate or change an outcome at somebody's expense, right?

[00:28:35] Those are the kinds of things, of course, that those ethical principles haven't changed ever, and it's just as true for AI as anything else.

[00:28:41] One of the unique challenges of AI is that it is a black box, so sometimes it's hard to know what decisions it's like or what methodology it's using to make a set of decisions.

[00:28:51] So it can have, for example, bias, right, to make a decision around supplier A to make its supply chain, you know, supplier A versus supplier B.

[00:29:00] And supplier B won the choice in the bid, but it was because there was some built-in bias in the algorithm.

[00:29:06] Yeah, or if the data, the historical data you have is all prior to March of 2020, and it changed 90 degrees after March of 2020, you're skewing the inputs there.

[00:29:21] Yes.

[00:29:21] Now, that's less of an even ethical question.

[00:29:23] That's more of just a data quality problem or, yes, you're actually just getting the wrong answer.

[00:29:29] That's why humans are important, too, right?

[00:29:31] Like, you need to spot check to make sure this is right.

[00:29:33] Exactly, right?

[00:29:34] So there's a quality piece of the challenge, of course.

[00:29:37] And there's a whole set of processes called machine learning operations or in the modern language, now large language model operations that are designed to sort of have an internal quality check process, plan, do, check, act around this.

[00:29:50] But still, that doesn't solve the problem of bias risk or manipulation because we're using AI to impersonate, right, a real person, all those types of secondary things.

[00:30:01] That nefarious actors could and will do, right?

[00:30:06] And so we set ourselves up from a policymaking perspective as well as a management perspective to handle those inevitable outcomes that are going to be there as long as humans are human.

[00:30:15] It's a big hairy monster when you talk about ethical because, you know, we all remember Milli Vanilli, right?

[00:30:21] We're lip syncing.

[00:30:22] But honestly, you know, auto-tunes, like auto-tunes, is that really their voice or is it not their voice?

[00:30:28] Well, I was just going to give the example of my 15-year-old son over the summer had a health class online and he was investigating, you know, the chat GPTs of the world.

[00:30:39] And so he put in his question and it gave him his answer.

[00:30:42] And my son copied and pasted the answer.

[00:30:45] And it said, you know, generated by chat GPT.

[00:30:49] Come on, buddy.

[00:30:49] If you're going to do it, that's transparency, I guess.

[00:30:52] Too much.

[00:30:53] Too much.

[00:30:54] Unintended transparency.

[00:30:55] Yes.

[00:30:56] Unintended transparency.

[00:30:57] Yeah.

[00:30:58] No, yeah.

[00:30:58] Teenagers will, that's our, like, our one built-in advantage as parents is that, you know, we kind of like, or maybe one step ahead on what it means to cheat.

[00:31:07] We can catch that.

[00:31:08] So true.

[00:31:09] Yes.

[00:31:09] And all the other words.

[00:31:11] I learned a new one last week.

[00:31:12] Reed, I'll call you later.

[00:31:13] Yeah.

[00:31:13] The teenage vocabulary, no cap.

[00:31:16] Yeah.

[00:31:17] So one last question before we kind of wrap it up.

[00:31:20] Crystal ball.

[00:31:21] Where do you see this kind of technology in five to 10 years?

[00:31:24] Or what impact in supply chain five, 10 years from now?

[00:31:29] Either or.

[00:31:29] I think there's going to be more and more of the Sheehan story, hopefully married to ethical practices.

[00:31:36] So it's not just about going faster and being more responsive, but also more lead and waste free.

[00:31:41] Right?

[00:31:42] That's a hopeful statement.

[00:31:43] But I think that my optimistic self says that that's going to happen as well.

[00:31:47] Because consumers will still vote with their feet when it comes to behaviors that are ethical as well as efficient.

[00:31:53] So the clock speed will go faster.

[00:31:56] The degree of prediction will go up.

[00:31:58] The closing of the loop with machines working with other machines will mean that we're going to have.

[00:32:05] Real IoT.

[00:32:06] Yeah.

[00:32:07] IoT, but also like it's an intelligent.

[00:32:10] It's not just connected, right?

[00:32:11] It's self-navigating, controlling systems.

[00:32:14] Hopefully with, you know, we've done the right things to put in governance mechanisms so that they're going the way we want them to.

[00:32:19] But autonomy, right, at a level that we just, that we think about when we think about self-driving cars, but apply it to supply chains, same thing.

[00:32:27] Yeah.

[00:32:27] It's amazing to me.

[00:32:28] I really feel fortunate and lucky to be alive during this time because we are going through a technology revolution, just like our forefathers went through an industrial revolution.

[00:32:39] There's no doubt about it.

[00:32:40] It's very exciting.

[00:32:41] All right.

[00:32:42] So we're going to wrap this up.

[00:32:43] And Steve, this has just been truly enlightening and thought-provoking.

[00:32:47] The last question we have for you, we ask of almost all of our listeners and not our listeners, but our guests for our listeners, what's your favorite technology that you're using today?

[00:32:58] Either personally or in your business life, either or doesn't matter, but what's your favorite technology?

[00:33:05] You're like, I use it every day or use it every week.

[00:33:08] I can't wait to use it.

[00:33:09] It just, what is it?

[00:33:10] Well, so I was going to say my bike, but that probably doesn't qualify.

[00:33:15] It's non-digital.

[00:33:17] Yeah.

[00:33:17] In the world of analog, yeah.

[00:33:19] I'm still fascinated with the efficiency of the bicycle.

[00:33:22] I mean, when I think about the breakthrough innovations, the ability to go from walking to going at that speed with two wheels in a chain, I'm still in awe of those fundamental foundational innovations like the bicycle.

[00:33:32] Having said that, in terms of cutting-edge technologies, I think I really do believe that, to your point about experimentation, that we should all be experimenting.

[00:33:40] I am.

[00:33:41] We are.

[00:33:42] With generative AI as a thought partner and as an assistant, I think we're still just discovering the potential of that tool.

[00:33:51] And the productivity gains are real, but also, frankly, the energy you get back from avoiding some of the mundane tasks of getting started on a 10-point list of ideas so that you can think about the implications of those 10 items on the list.

[00:34:08] That freedom to think bigger, it comes from removing some of my mundane thinking is the power of GNI for me personally, and it's super exciting.

[00:34:17] Well, and you said something that struck me.

[00:34:19] It makes you better at what you do.

[00:34:22] It makes you better at what you do to help that productivity and thinking.

[00:34:26] Exactly.

[00:34:27] This has been fantastic, Steve.

[00:34:28] We can't thank you enough for spending the time with us, and we look forward to talking to you again in the near future.

[00:34:35] Likewise.

[00:34:35] Pleasure.

[00:34:36] Thanks for having me on.

[00:34:38] Thank you for joining the Next Level Supply Chain with GS1US.

[00:34:41] If you enjoyed today's show, you can subscribe to our feed or explore more great episodes wherever you get your podcasts.

[00:34:48] Don't forget to share and follow us on social media.

[00:34:51] Thanks again, and we'll see you next time.