EP 596: Warehouse Orchestration with AutoScheduler
21:39 Share

Spoiler alert, your team hates manual unloading. And who can blame them? It's slow, risky, and exhausting. Tawi's Container Unloader eliminates the pain.

Lifting loads safely and efficiently from inside containers. Engineered to lift heavy goods ergonomically from deep inside containers, it accelerates unloading while keeping your team safe from strain. Safer workers, faster turnaround, and better efficiency. Better for your people, better for your bottom line. Lift smarter. Visit Tawi.com today. That's T-A-W-I dot com.

The New Warehouse podcast hosted by Kevin Lawton is your source for insights and ideas from the distribution, transportation, and logistics industry. A new episode every Monday morning brings you the latest from industry experts and thought leaders. And now, here's Kevin.

We are here at ProMat 2025 and joining me here on our lovely couch this year is Keith Moore, CEO over at Auto Scheduler AI. Keith.

How are you? I'm doing great. Thank you so much for having me. Definitely, definitely. I know you're happy to be sitting down for a little bit here, right? It is, yeah. It's been a long day, right? I was saying before the show, like, didn't have a chance to eat anything today. I've literally been talking all day. So this is nice. I appreciate you having me. Definitely, of course. Always good to get you on. We've had you on a couple times, right? Yeah.

early on pretty much in the podcast we had you on. So it's been great to see your journey along with our journey, right? You're right. That was like right after we launched. Yeah. And so I'd like to credit the company's success back to the program, obviously. Yeah, definitely. I mean, we were certainly a catalyst for the launch. Yeah.

It's been valuable for us. So I'm glad that we've had a long-term relationship. Yeah, yeah. So it's great to see you. Great to catch up with you here a little bit. So kick us off here. Tell us, for people that are not familiar, give us a little kind of brief overview of AutoScheduler. Yeah, so AutoScheduler is focused on warehouse orchestration, which is a new term. You know, five years ago when we launched the business, nobody knew what it was. And now there's actually roles created for orchestration in warehouses today.

Interesting. Yes. So what we focus on, it's really three separate things. So the first piece is data harmonization across warehouses. A lot of the facilities we operate in, they're not just a four-wall distribution facility, which we operate in, but also a plant warehouse. So you're looking at all this complexity where you have a manufacturing schedule, you have a yard management system, your inventory management system, your WMS, some automation in a WES or WCS system.

We do all the data harmonization. So bring it into one place where anybody can view anything they care about the facility in a single location. The second piece is then advanced optimization. So we leverage artificial intelligence to plan out what is the work that needs to be done? What are the flows that work is going to go through and what is the right way to flow all of the inventory, whether to production through production off production or through the facility for order management in bounds in as well. Uh,

what person is doing what work at what time in what location for the whole facility. So we're doing that mathematics and that's a continuous dynamic process. Whereas conditions change, we're replanning and telling you what you need to do next, not just for the next shift, but for the next day to two days. And then the third piece of what we're doing is we're harnessing, then capturing all of that data on,

What did we say to do? What was the optimal output of how you could have run the facility? What did you do? And where's the opportunity for gained margin? So you ended up traveling 80 miles more than you could have if you had used better door choices based on inventory locations in the building. You had...

heavy utilization of labor in shift one, but looking back at shift two, we thought you could have gotten a lot more work done, or you could have actually done early voluntary call-offs and told some of your work to go home because you were far enough ahead as a facility. Yeah. Right? So that third piece is capturing and getting all the way down to what is your optimal cost per unit and what are you actually achieving based on plan to optimal and

and attainment. Interesting, interesting. So kind of creating that orchestration to ultimately lead to the, I guess, ultimate optimization, right? That's exactly it. So, and you said something interesting there because I don't think, the first time we talked, if I recall, I don't remember you mentioning anything about actual roles being created around orchestration now. So that sounds a little bit like an evolution there. So,

Talk to us a little bit about that and what those roles look like. I would love to say that was us, right? Because I've had a Google alert for warehouse orchestration since we started the company.

And it was only ever about us until about a year ago where you started to see other people talk about it. But now, more often than not, a lot of folks that are engaging with our business, you have your continuous improvement, your industrial engineering teams and warehouses, but now people are creating these orchestration roles because they've realized you have people managing a facility and those people now have to share data across five different systems to

to run the site. And so continuous improvement with like a technology angle is moving into orchestration and what orchestration looks like in distribution.

Interesting, interesting. So, I mean, I would say that that role should be called the warehouse conductor, right? I mean, that would make sense. That would make a lot of sense. Or the maestro. The maestro, yeah. I'm a warehouse maestro. Thank you. Yes, thank you very much. Yes. So, very interesting there. And obviously, you know, a lot of touch points there that you're hitting within the operation, right, to be able to make that full optimization. But we're here at ProMac, right? So, we come to ProMac because we want to know what's the latest with everything.

So tell us kind of what's the latest with AutoScheduler, what's new with AutoScheduler?

Yeah, so I covered a little bit of it, but the world's been changing a lot. Yeah, I think it's an understatement. Especially in distribution. I love it, right? The past five years in distribution and warehousing has just been fantastic. Plug for the show is if you listen to Kevin's show, he does a really good job of actually staying up to date on all the trends, what we're seeing. But from my perspective, what that means is there's a few things that are kind of coming to a head is...

this labor problem that happened over COVID didn't go away. We've just stopped focusing on it. People have gotten to the point of maturity in automation projects. So a lot of people started really investing in automation over COVID. So 21, 22, 23. They've hit the point where a lot of those projects are now completed, but they're not realizing the return on investment they were expecting. So they were expecting, hey,

our availability of labor is going down. We need to offset that. We're going to invest in automation. So they invest in automation. They're not seeing the return they might've hoped for. And that coupled with

more demand. So, so when I say more demand, what I mean is complexity of orders has changed significantly over five years where, you know, it used to be a lot of operations were like pallet in pallet out and CPG and food and beverage mix of products has changed. People are operating, you know, so Frito operates a lot of their own like fulfillment operations. Right.

and customer of ours, we're familiar with, and this is public knowledge. And so like the order profiles that people have to deliver on inside of these facilities is changing as well. And that's kind of been the perfect storm for orchestration technology.

because you have this automation that is really hard to manage, that isn't really netting the ROI people had hoped for. You need to better manage the labor that you do have in tandem with the automation that you've put in that handles a particular set of flows. And then the flows you've introduced in the building, it's no longer just grab something off a truck and put something on a truck.

You have hundreds of different unique ways products can flow through every facility and you have to manage all of that in concert as things are constantly breaking in a facility. Yeah, yeah, absolutely. And I think that's such a great point. I mean, things are constantly breaking in a facility, right? And I remember my days as an ops manager and it was always something. I mean, there's always something there that's like not going as you expected. You come in the day like this is the plan for today, right?

And within two hours, it's totally different plan. It's totally different plan, right? So having something to be able to create that optimization and kind of align those things a little better, I think is definitely key. And I could see how as those order profiles, like you said, continue to expand or change or have more variations to them, right? It becomes more complex to try and plan around that and try and make sure everything is lining up and being as effectively utilized as possible.

So very interesting there. So you mentioned that you guys leverage AI as well, right? And there's a lot of AI hype out here, right? You know, I think, I would say I haven't seen as much AI here at ProMap, but certainly other conferences have.

you know, there's a lot of AI, right, that are tagged on to these names, right, including yours, I guess. Certainly. So tell us a little bit about, I guess, your view on the AI hype and, you know, how it's kind of

falling in and being utilized by auto-scheduler. Yeah. So the first thing I want to do it like level set on definition of AI. Let's do that. Which is the ability to replicate human activity in automated mathematics, basically. So then to break that down a little bit further, like the AI that's extremely popular these days is a subset called generative AI, which is going to be, I mean, I think everybody knows chat GPT and Gemini from Google. Yeah. Right. But,

But AI is much more than just generative AI. One of the big challenges, and we're at a warehousing conference right now, which I love, right? This is the Super Bowl for people like us. Yeah, yeah. This is the spot. And so...

generative AI really struggles in a warehouse because warehouses have a particular set of data that they let right so your inventory management systems your task management systems your yard management systems like generative AI works because it was trained on a massive corpus of data which is just a body of data and it

And it is extremely focused at either generating text or generating images, but it's not trained on the semi-structured data that we generally deal with inside of warehouses. So it's very good at supplementing other types of AI and machine learning inside of warehouses. But it is, like, I get the question all the time,

Will generative AI run a warehouse in the next few months? And the answer is ubiquitously no. Okay. Because it doesn't understand the flows and constraints to go through a warehouse. But a generative AI can take other AI and machine learning models, like what we're leveraging at Autoscheduler, which are doing all sorts of predictions about what the future is going to hold. They're doing then all sorts of decision optimization models

to say, given all of the different ways we can run this warehouse, there's millions of different ways, there's actually more ways to run a warehouse on any given day than there are atoms in the universe. What do we do? Really? Yeah, yeah. So that's like a fun combinatorial optimization fact. So what do we do? So that is most of the AI that is in use today. And it is either to predict and inform what is going to happen, what should happen. So will a truck show up on time based on carrier, based on everything else?

And then will a person show up potentially? And then it's what decisions should be made given all of the needs of the building and the trade-offs that could be made to optimize overall throughput, minimize cost. So that's the best fit for AI in distribution today. There's a bunch of unique vertical places that we could... And I'm happy to provide examples. I just don't want to talk for too long.

Now give us a couple examples. All right. So obviously what we do with orchestration, we're doing the whole flow of who's doing what, when are they doing it, where are they doing it. That's AI. There's AI uses for solving the traveling salesman problem. So if you need to better optimize routes for picking where you have a cart, you only have so much space on that cart.

And you will have so many orders you need to pick that are all due at different times. Like, what is the best way to wave those orders out? That's a great use for AI machine learning and optimization. Slotting is another good one. How do I lay out my facility based on seasonality, dynamicism of orders? Like, you know, fast movers, slow movers, ABC goods. There's also really cool uses of AI and vision systems for...

Truck management and yard management. Some for inventory management. I think those are still working through maturity and getting perfect. But the yard management stuff is a home run. Yeah. Interesting, interesting. Yeah, and I think it's really interesting because there is... It's such a broad...

kind of word or, I guess, acronym that's thrown around, AI, right? Is it like, oh, this is AI. It's a marketing word. That's AI, right, yeah. And, you know, it's interesting to hear you kind of break it down like that a little bit. And so now if we look at that, you know, I think there's a lot of question, you know, I mean, a lot of the solutions we see here, like whether it's AI or robotics or automation, I think there's a lot of question for a lot of those, you know, end users, operators, or like how do I even...

How do I even get started? So from an AI perspective, I mean, how do you kind of help, whether it's your customers or just people in general, help them to understand how do I get started on getting AI into my operation and where do I start? Yeah. I'll give you the academic answer because like the Keith is a CEO of a company that does it. Answer is like, yeah, just work with me. Life will be easy.

The academic answer is there's two things you have to do before you think about machine learning and AI projects. The first one is if you put a person in a room with all of the data that existed that you were planning on throwing. So you have a problem, you have data. If I put a person in a room with that problem and that data and an infinite amount of time, could they solve it?

And if the answer is no, it's not a good AI project. Because that's the real value is the concept of AI is you get intellect at scale and consistently, repeatedly continue it. All those things. So that's the first question to ask yourself. The second question to ask yourself is, can I map? So do I have data?

Garbage in, garbage out. That's certainly a standard. But the piece that people miss after they get through the garbage in, garbage out trope is if you can't map your data to a process. So not only do I know the process inside of my facility, but I can show you exactly where in my data that process manifests.

You'll certainly fall flat. So doing that, that process mapping is one of the first things that I would recommend people do. At which point when they can do that, then it's a great way to take on machine learning and AI because they've got all the bones in place to start to do algorithmic development.

Interesting, interesting. Yeah. Yeah, I mean, I think it's such a great point there. And you talked about the garbage in, garbage out, right? Yeah. And you also said that you guys do data harmonization, right? Yeah. So tell us a little bit about from that front because what I hear in a lot of scenarios is that

you know, companies are like, okay, yeah, let's go after the solution, whether it's, whether it's AI or robotics or something like that. And then all of a sudden there's kind of a, a shock when they realize like, Oh, their data is like not so clean. It's not so great. And now there's this whole other project that's come up to, to try and clean that up. So, so tell us a little bit about kind of the,

the importance of that and how do you navigate that with some customers or potential customers? Yeah. So navigating is, it's always a challenge and how people are set up is different. There's it restrictions that certainly make life interesting or challenging. Uh, but,

Maybe I'll tell a story, which is a good example of this. I was at a site the other day, and it was a manufacturing facility where most manufacturing facilities have raw materials and then finished goods out the other side. So your finished goods you're shipping downstream either to deploy to your own sites or to customers. Raw materials, your job is to support production. So when you ask, okay, well, how do you run the site? Well, we run the site. We have a group that we pay to basically run both sides of the operation. Okay, well...

Like WMS-wise, what are you using? Oh, we have a raw materials WMS and a finished goods WMS. So same site, same people doing the work, two completely different systems that manage the processes. And so part of what you have to do as somebody who wants to bring AI to that operation is, okay, how do we take these two systems and marry them so it is one pane of glass that you need to look at? And that's certainly the first step.

Companies that five years ago invested in data lakes or systems that could harmonize data for them, they are so much further ahead at this point on the AI journey because they did that. And where you haven't done it, that's part of my job is how do I make things easy, modular, and harmonized for these customers so that they don't have to go invest all that time and effort to do that. Yeah, exactly. And I think it's super interesting because...

you know, like in my career as like ops manager, like I specifically remember, you know, having a couple direct reports and stuff where, you know,

They're like, ah, like I hate data, right? I hate Excel. Like don't want to work with it and all this stuff. And, you know, it's like stressing out. So even though we were not doing anything like that was super high tech, like I'm still like, ah, but that data can do so much for us. Like we've got to get it right. And I think that, you know, like you just said, you know, there's people that invested in, you know, five years ago, like they're in a much better spot now. So how would you say, you know, for somebody that's, you know, coming in here to ProMap where we're at right now and,

they're looking at so many solutions and they're thinking like, okay, we need to modernize our operations. We need to get to the next level. What would you say is like the first step for them? Modernizing your operations. Like everybody should do an analysis of business outcomes they want to achieve. Like, you know, a WMS upgrade is like open heart surgery. It's like best case scenario. You keep living. And so, yeah,

like that. Yeah. I stole that from somebody else. I wish I could claim that as my own, but modernizing for modernization sake is not the right thing to do. There has to be a set of business objectives you want to achieve. Technology is an enabler. And, you know, obviously I should be touting like everybody needs AI, everybody needs auto scheduler. But at the end of the day, my job is to help people do their job more effectively and more easily.

And what is more effective and what is, you know, more easily. Yeah. That really depends on what outcomes you want from the business, whether you're trying to squeeze margin from facilities, whether you're trying to maximize throughput, how you're planning and what your strategy and your supply chain actually is. Like, that's always step one. People, process, technology. Like, we

Like we're one of the three pillars, but not by any means the whole thing. Yeah, yeah, absolutely. And I think it's really, really great to think about it that way. And, you know, I've heard some views on that where, you know, you come and you want to stay away from the shiny objects in a sense, right? Like think about, you know, what actually is going to solve the problems that I have, right? And it's kind of what you're saying through, you know, getting those business goals in line and really understanding like, well, what am I trying to accomplish at the end of the day, right? Exactly.

So very interesting with you here, Keith, and I really appreciate you coming by to the booth. We're happy we give you a seat for a little bit. I know you're working hard this week. Sincerely appreciate it. Definitely, definitely. So always great to catch up with you. If people are interested in learning more about AutoScheduler, getting in touch, what's the best way to do that? Yeah, so check us out, autoscheduler.ai.

It's the website. We have all sorts of forms. We actually have a really fun game on people can actually play. Can you run a warehouse successfully? And we have designed the game so that you will always lose points. So it's actually really fun. It turns out it's super simple. It's like a three-door warehouse with just a few shipments and receipts and

We ended up basically using our technology to solve, like, what is the optimal just so we can go figure out who actually did the best. I don't think anybody's ever gotten a perfect score yet. That sounds like a challenge to me.

But, you know, check us out on autoschedule.ai or, you know, you can add me on LinkedIn. I try to produce and promote educational content because I'm very passionate about AI, very passionate about the space we play. Yeah, absolutely. So we'll definitely put all that information at thenewwarehouse.com and also in the show notes here so people can easily find them. So, Keith, thank you once again for joining me here at ProMat 2025.

You've been listening to the New Warehouse Podcast with Kevin Lawton. Subscribe and check us out online at thenewwarehouse.com. Enjoyed this episode? Make sure you are subscribed to the podcast and for more content from The New Warehouse, find us on LinkedIn and YouTube. Links to subscribe can be found in the show notes. And for everything The New Warehouse, head to thenewwarehouse.com.