Why AI Rollouts Stall (And How to Start Anyway)

Alvaro Cuba

Guys in manufacturing, we have more than a half a million open jobs in the U.S. So we need more people, not less. And we need people with better, different skills. So we need to train the people more. What that tells you is technology is not a threat in manufacturing, supply chain, in procurement. Bring your people with you. Help them, train them, and we need more people.

 

Ed Ballina

Hi, I’m Ed Ballina.

 

Alvaro Cuba

Hello guys, Alvaro Cuba here.

 

Ed Ballina

Well, welcome to the Manufacturing Meetup Podcast, the show where we kick back on our downtime and we tell great stories, much of which are true, but we actually learn from each other and share our little tips and tricks and love to get them from you. And we’re also known for wearing silly hats like that.

 

Alvaro Cuba

I can see that summer is coming.

 

Ed Ballina

Summer, yes. So my hat, I was fishing or attempting to fish in Canada last week for five days. Note to self, never go up there until the end of May, because the weather was horrible. I caught a bluegill about this big and one other fish about this big. But the time was phenomenal. I learned a lot about the Gananoque area in the St. Lawrence River and it was blessed. So this was my fishing hat. That’s why I’m wearing it today.

 

Alvaro Cuba

How about your hat? I know. This is just my home one, Miami Beach, because next couple months I am put in town just for the soccer tournament. So no traveling for me in the next two months.

 

Ed Ballina

You are grounded, my friend.

 

Alvaro Cuba

I’m grounded for two months, but worth it. I can tell you guys, worth it.

 

Ed Ballina

You know, that passion, that soccer passion is incredible. You know, those of us have had an opportunity to kind of visit and live in other countries in the world. I mean, as crazy as we are about our sports here, the soccer fans are like one level above that almost. And it’s great, great to watch. So anyway, that’s the story with the hat, right? And hopefully I’ll get to take it off because somebody said I look like a little old lady in the play or something.

 

Alvaro Cuba

Amigo, I haven’t said anything.

 

Ed Ballina

Okay. It was a great show. So Alvaro recently sent an article to us about P&G supply chain and kind of where they’re going. And it really inspired today’s topic, right? So please listen to our reaction to this and a little bit of an expansion and tell our meetup buddies and also kind of let us know why did it catch your attention? We’ll tell you what caught ours. Alvaro.

 

Alvaro Cuba

Yeah. In reality, we are out in the plants and we are seeing so much great, cool technology in safety, in quality, with AI, in maintenance and doing amazing things. But what we are not seeing is connecting all these and bringing the plants to the next level. So we were saying, what is missing? What is not connecting? If the technology is there, why we are not seeing this step up in driving excellence? But first, please hit the subscribe button at the bottom of the screen so you can follow our conversations and stay with us, give us your comments. And with that, let’s get started.

 

Ed Ballina

Let’s do it.

 

Alvaro Cuba

So let me set the stage first with a couple of examples on CPGs and as we were saying in the introduction about the capabilities. We saw an article on P&G that was explaining a lot of what these guys are doing. And basically they are working in two areas. One is manufacturing and warehousing. They are doing a lot of automation. And even they are starting to pilot in nine different plants lights out, which means four hours at night, nobody in the plant and the plant runs alone. And they are finding amazing, between 15 and 60% increase in productivity in these cases.

 

And on the other side, they are working on integrating systems. Even P&G, with all the experience, and Ed can tell us a little bit more, he worked there trying to integrate customer order, production planning, material order. And with this, they are projecting $1.5 billion in revenue on cutting costs, take service to 98% on-shelf availability and those kinds of numbers. We also have seen Pepsi going AI and twin technologies. Hershey just announced a hundred million dollars in inventory reductions because of AI technology. So as Ed and I always say, the ideal is you need to hit revenue, cost and inventory. But what we are seeing is pockets of it versus big integrations, global rollouts, connecting all the dots. What are you seeing out there Ed?

 

Ed Ballina

Yeah, spot on. And we had also looked at some of the McKinsey data. McKinsey had an interesting report that just came out, and we’ll share the link in the show notes, folks, so you can access it yourself. They’re quick reads, but they’re full of really great information. So here are the highlights. Over a hundred manufacturing COOs were surveyed, and not surprisingly, seventy-four percent of them said yes, of course we have a global production system. And what are we doing from an AI standpoint? Well, we have AI, but we’re only about twenty-nine or thirty percent implemented across all sites. And nobody, not surprisingly, said that we’re fully integrated. So to Alvaro’s point, we’ve got pockets. We have pockets of the same agents working in facilities, right? Sometimes different agents trying to do the same thing. So the governance of this is starting to get a little crazy, but nobody has gotten to the point where we are standardizing and then integrating both vertically and horizontally, right? Because you want to integrate all the way up and down the supply chain. You also want to integrate with your customers, marketing, et cetera.

 

Alvaro Cuba

And it happened to us. And it happened to us. I’m sure you remember the times of SAP integration. We got the system and SAP was doing everything from one end to the other. And it was the holy grail. Trying to get it to work integrated across, or Lean or Six Sigma, always pockets and things like that. And I’m sure you had all the same experience.

 

Ed Ballina

Yeah. Some of the things I saw work is when I remember the early days at P&G Mehoopany we decided to roll out statistical process control and Deming TQA, right? And there was no question. The training went on, we brought in control charting and all that, and we were all moving together. And that had a tremendous, tremendous value. That’s just a small example. One that I’d like to talk about that highlights what Alvaro is saying is CMMS. So CMMS is an example where we’ve instituted it in a bunch of places and it hasn’t quite delivered. Why do good systems stall out? CMMS systems. Most companies have invested enormous amounts of money in computerized maintenance management systems. That’s CMMS. Or some sort of a system that allows you to track preventive maintenance, equipment history. The real good ones have the potential to link to storeroom parts and even order parts for you when they’re pulled out of inventory through a work order process, right? This is not scientific, but I would tell you at least 50% of the places that I have had experience with instituted CMMS systems and have maybe about a 50% take rate.

 

The PM work orders, half of the time, they’re not really scrubbed to be effective. We don’t capture a lot of the reactive or breakdown work, and we don’t build that into the equipment histories to reach the promise of CMMS, which is, hey, you have all this data, we’re gonna tell you what equipment fails first, and here’s the cost and all that. So the promise was never fully delivered. And what I saw at P&G, what worked was the culture and the discipline, right? When P&G said we’re going in this direction, once it got there, there was no, well, I don’t want to do that or I want to try my own homegrown version. No, this is the path. And a huge amount of discipline around that. There were no side systems. SAP, Alvaro, how many times do you roll out SAP and find out you have 170 processes that nobody ever identified, but they run your enterprise out on Excel spreadsheets?

 

Alvaro Cuba

And what was the percentage of SAP that even the best companies use? 20, 30%, 40%?

 

Ed Ballina

Right. And some of this kind of makes sense, folks, because SAP strives to be a system-wide enterprise, right? They have modules for finance, for HR, for manufacturing, for logistics, everything under the sun. But that old adage, the journeyman of all trades but master of none. By definition, when you try to grab a whole enterprise, there are some modules in SAP that are phenomenal and there are some that are not so great. I will use the example of the predictive models compared to Augury, not even close, right? So you always wind up in this back and forth. But if I kind of roll it up, implementation, you have to have the right culture, you have to have the right discipline. And the last thing that I will add is you need patience. Procter and Gamble, at least in the time I was with them, understood that these journeys were three to ten years. And they didn’t just fund it for two years and then say, oops, it didn’t deliver instant pudding and cut the program. I’ve been in some enterprises where that has happened and we have relaunched some of these programs three or four times under different monikers. That’s not the way to do it.

 

Alvaro Cuba

But I would say that there is light at the end of the tunnel. Technology is a great enabler these days. Ed was mentioning the McKinsey study with 100 COOs. And what they have found is that now that the technology, AI and the technology is there, it is less about uniformity of tools, less about standardization across, because that doesn’t help to turn insights into execution. What they are finding is that the key is how fast you get data insights. That’s the first element they say.

 

And getting the right data, we talk about this, it’s transparency, gives you a baseline, benchmark, visibility. And that’s the starting point. Once you get that, what they are finding is the second element is referring to what Ed was saying. It’s culture, it’s people. It deals with capability, training of the people, flexibility, repeatability of behavior. And even they throw one in which is coaching, which I found really, really interesting. And if, just to finish the idea, if you have this data and you have the people with enough flexibility and power, and the data is telling you what to do, you can react very quickly. And then technology helps you to accelerate the results. Then you start getting the power of all this technology.

 

Ed Ballina

Your point around standardization is an interesting one, right? Because at a base level, you need standardization, right? If you’re building a house, you need your foundation to have all the blocks be at the same level and all that. No different in these enterprises. So you need to standardize on how you measure performance, right? What do I call efficiency? What does that mean? Do I use OEE or some other method?

 

I was in a situation where we would acquire smaller companies and the first thing that I usually did was I standardized their output expectations because some of these outfits were holding themselves accountable, let’s say, for 400 bottles a minute on a filler that we knew our OEM standards were 600, right? So you’ve got to level set. But once you get that level, you have to allow flexibility of solutions. Back to the SAP discussion, right? They try a one-size-fits-all approach, and that’s not necessarily the right model anymore. It’s about having the flexibility. If you think about agents, right? These agents, you have a predictive maintenance agent that pops up and gives you data about vibration and all this. And it outputs that in a certain language. You have another one that does supply chain planning. You need those to be linked because you’re going to shut equipment down. But that speaks in a different language. You need an ontology above that that translates it and provides people like Alvaro and I with insights. So we don’t have to be the translator, because it’s impossible.

 

Alvaro Cuba

But there is a transformation. It’s the next wave, industry 4.0, 5.0. It’s the next wave. So we are hopeful that we don’t need to deal with the SAPs or the Leans or the TPMs the way we needed them. We need the base, the structure, we need the systems, but we don’t need to deal with them in the old way. Now that we have AI technology, think about ChatGPT. In the old way, you were going and spending two or three hours in the library or looking, digging on the internet. You had massive amounts of data and then you had to compress it and try to get the nuggets that you needed and then interpret that in a way that people can understand. Now you just ask ChatGPT and all that took you a lot of time. Even if you could, because in many cases you couldn’t, now it gives you that. So it’s clear that the world is changing and the technology is exploding. So the question here, and this moves us to segment three, is how using that technology, we in the companies use that to accelerate this improvement and really start harnessing the power of this technology and getting that into the results.

 

Ed Ballina

Yeah, it really is amazing. And Alvaro and Sara, who’s our great producer here, have heard a lot about me with ChatGPT, where I went from zero to a little crazy these days fast. Here’s an example. I had a problem getting a side view mirror for my vehicle replaced. I’m not gonna mention names, but it took that company seven months to find that mirror in the time that my vehicle was not fully operable. Well, I’ve gone back and forth, and one question I thought to myself was, how unique is this mirror? I’m driving down the road and I said, Hey Siri, ask ChatGPT, and it came on and I described what I was asking for. It said that mirror is also available for this vehicle and this vehicle and that vehicle. So I’m like, when I’m talking to them, I don’t want to hear that mine is a bespoke, carved away in Regensburg, Germany. No, no, you put this in five other year models. So don’t tell me this thing was as rare as hen’s teeth. I will stop with that, but I thought you would appreciate that one.

 

So what does this mean if you’re starting a journey? Look, it’s not easy, right? You can’t automatically compare yourselves against a P&G or some of these other Fortune 50 leaders in the field. Having spent some time at a Procter and Gamble, and to this day they remain a company that I hugely admire, they have had decades of process knowledge, adherence, discipline, culture, right? It’s a lot easier when your data and all that is fairly organized to be able to start doing some of this than if you’ve got to hunt around and get people with the discipline of even capturing information. So the funny thing is the technology is moving so fast that by the time P&G launches this Supply Chain 3.0, which by the way, as you heard, is mostly around automation, right? They’re going to be behind, not behind, but kind of behind where the technology potential is. But that’s okay because they will still be leading the field. So if you’re a mid-sized manufacturer and you don’t have a perfect operating system, where does that leave you and how do you just get started? And before I turn over to Alvaro, I would like to note that I see a pattern here. I’m the bad guy that sets up the problem. Then Alvaro comes in as the savior to say, Ed said this, let me show you how you deal with that. Sorry.

 

Alvaro Cuba

Okay. We are part of a team, right? I’ll take it. I’ll take it. We’ll turn tables. We always do. Think about the way I can think about it. Ed was talking about the P&Gs, the Toyotas, the Motorolas, for instance, how long it took them to get to a three-layer plant. How much organization and things had to happen to get to that level.

 

Ed Ballina

Training alone, right? Massive.

 

Alvaro Cuba

Training alone, systems, interactions, and many things. Fast forward, technology today. So even if you are a mid-size company, AI takes care of the insights. Ed was talking about the maintenance example. But for safety, you put some simple cameras. Or for quality control, you put some simple cameras, you connect those into wifi, they start sending information. And now you know what AI can do with that information, the same as ChatGPT. So you bypass a lot of that and that information even translates into actions. And those actions go into your people.

 

That comes the second element. You have to train and empower your people. That’s the second element and it’s squarely on you. With those two elements, like McKinsey was saying, if I have good insights and I have people trained, flexible and ready to go, you launch into action. So do it for predictive maintenance or do it for safety or do it for quality control. Just jump into one of them. Very, very easy, very fast. And then learn, take the learning cycle, learn what works, what doesn’t work. You are not investing a lot of money. If you are failing, fail fast, not much money, just come back and do it the other way. But there is a lot that today the technology can do that allows you to accelerate the paths to get to where it took 10 years at P&G, or at Kraft Foods. Even the Japanese, it took 20 years. In these days, when you see the new companies, when you see what Apple is doing with OpenAI, they are doing it in six months. They are integrating completely OpenAI into the Apple system in six months. Why? Because they are using the technology to help them do that. What do you think, Ed?

 

Ed Ballina

Yeah, the barrier today is much lower than we used to think, right? I remember, you guys know I’m a diehard infrared thermography person. The first camera I ever saw was the size of a Betamax, about this big, and cost sixty-five thousand dollars and was cooled by helium gas. Now they’re like this and they’re like four hundred bucks, you can attach them to your cell phone. So the technology has lowered the price of entry. And I’ll give you a quick story. When I was still involved full time in the beverage business, we installed equipment sensors on about 12 pieces of equipment in our Fresno facility. And within about three months, so not a huge investment, very reasonable, there were critical pieces of equipment, our water pumps, et cetera. Within three months, they identified issues that would shut the whole plant down for eight hours each, if not longer. And the rate of return on that was astronomical. This is not that expensive. The technology is available.

 

And Alvaro had talked a little bit about the use of digital twins. I’ve mentioned this before. I was at a session where I was able to see a digital twin at Pepsi’s R&D labs and what that can do for a line, and I was blown away. I mean, literally you have a model of a bottle line and you can load in some new bottle design that the marketing guys want, and run it on the line virtually at speed. I was watching, it was so cool. When it started at three hundred bottles a minute, the bottles looked pretty good. So as we hit about six hundred, they were wobbling and falling all over the place. We went back and told the marketing guy, no triangle points on the bottom. Sorry, no pyramids, doesn’t work, guy. I’m exaggerating, but sometimes you have to have the actual proof. So don’t be afraid of the technology and you can take it one bit at a time. You don’t have to conquer everything in one day. Start small, show the benefit, show the ROI, and you build credibility with the organization.

 

Alvaro Cuba

I think McKinsey, I love the way they paraphrase it. They say the power of continuous and connected insights into execution. With all we talk about, the continuous and connected insights, it’s about technology these days. Many years ago, it was about many other things, but today it’s about technology and it’s cheap, easy to implement, low barrier of entry. So you can go for it. And then the next piece is the people, the culture, the training, the skills, how the peer exchange works, how the leaders help to see the big picture and orchestrate it, governance and culture. So some key takeaways, Ed.

 

Ed Ballina

Yeah, listen, just take a step, okay? Just take a step down this road. I’m not gonna guarantee that you’re gonna get it perfect. Your first AI pilot may not go very well, but you will learn from it. So typically, guys like Alvaro and I, we’re asked to put together productivity plans and they include a big capital list that we know is gonna get cut in half, right? So you put some sacrificial lambs in there ahead of time. But you’re talking like an investment of, you may have 200 million dollars worth of capital to spend in a year. You’re investing for the next 20, 30 years, right? You’re gonna be paying depreciation on that equipment for that long. This stuff we’re talking about literally is in the tens of thousands of dollars in some cases and have returns that are in the months, not the years. So it’s easier to take a risk and not feel like you’re gonna get burnt. Also, know where you stand today. Be blatantly honest. If your baby’s ugly, call your baby ugly and then figure out, from the standpoint of building the strategy, where do you start? Alvaro gave you some of these. Start standardizing some of your measurements, start talking to your people about AI. Do a little research. There’s a ton of companies out there that provide agentic solutions. And for me right now, the money’s in agentic. Eventually it’ll be system optimization. That, in my opinion, is years from now.

 

Alvaro Cuba

Yeah. So I would say just a couple of things. Don’t be afraid. Go and get that technology. And then focus on your people. Flexible, curious people, educate them to learn. Nowadays with technology it’s easy to learn, but they need to have this habit to learn. And the last thing that I would say is there is a lot of buzz out there about technology and the people. Guys, in manufacturing we have more than a half a million open jobs in the U.S. So we need more people, not less. And we need people with better, different skills. So we need to train the people more. What that tells you is technology is not a threat in manufacturing, supply chain, in procurement. Bring your people with you, help them, train them, and we need more people. At the end, everyone is going to tell you technology coupled with the people is where you should go for a step up.

 

Ed Ballina

That’s the secret sauce right there. And look, folks, the current pushback that you’ve been hearing about AI is no different than every other industrial revolution we have experienced, right? I remember reading some novels and books where when the internal combustion engine first came out and there were people driving cars past horse and buggies, like it was gonna be the end of the world. My God, this is not gonna work, the horses are gonna get run away and they’re gonna kill people and all this. And every change, right? When we went from the shoemaker to the shoemaking factory, people thought, my god, that’s the end of labor. It’s never the end of labor. But we really do have to address this issue because there is a buzz out there that is creating a real negative wave about AI. They’re trying to tell people that means everybody’s gonna lose their job. No, it’s not. But we do have a social responsibility to craft a vision that enables people to see their place in that future society. And the good news is we’re looking to upskill your work, right? To make your work more interesting and less boring and utilize all of your talents.

 

Alvaro Cuba

With that, we are optimistic about this and we think it’s going to be a tremendous help and a tremendous source of growth for the people and for the companies. Because once you put together technology and the people, you are going to have a big step up. With that, folks, we come to the end of another episode. Controversial this one. But thank you so much for joining us at the Manufacturing Meetup. Really appreciate it. If you enjoyed, please follow or subscribe. Like us if you are watching on YouTube or leave us a review if you are listening on iTunes and please tell your friends. We love to hear from you and we love for you to be part of the show.

 

Ed Ballina

That’s awesome. Hey, here’s a thought. If you have a particular guest speaker that you’d like to have on the show, or somebody you’d like us to interview, hey, drop us a note, let us know. We’re always looking to bring new talent, new faces here to MMU and look for your suggestions. So if you like this and you want to keep the conversation going, you can email us at mmu@augury.com. We’ll also have the links in the show notes for this episode and we’ll see you next time. Happy Memorial Day all.

 

Alvaro Cuba

See you guys.