Podcast: How AI Is reshaping Lean manufacturing and continuous improvement

JJ: Yeah. In the realm of lean manufacturing, based on your observations. Where do you see AI at the moment in terms of, is it in its infancy? Some people are putting their toes in the water or, you know, what are you seeing there?

EL: Yeah, it's everybody is trying to do something with it. So it's very trendy. So all the client companies I've worked with when I was working in strategy sessions with them and working on strategy deployment and their Hoshin Conry, they're all saying, how can we better use this? How can we use AI to do different things? So everybody is experimenting. And that's the way I think you have to look at it is These are, pilot experiments. you look at some of the big technology companies and the rollouts from Google and Microsoft, they're putting all of this in for about agentic AI and different things that are going to help us. What I would tell people is, you know, you still have to have an overall operating system. And I think that Lean and the principles of Lean, continuous improvement, process-based thinking, get the workers involved, have respect for the people, and make the process be focused on the customer, is the right framework and operating system to help enable AI to flourish. And I just don't see it as AI replacing it. You know, when I see some of the leaders that are in the forefront of some of the thinking, they are, they're saying how can they better synthesize customer data? How can they use AI to help them aggregate, pull data from the data lakes that were out there and then actually make sense of the data? But you still have to have a person to interpret those results. And you also have to have a person who is understanding what are the inputs, because there is bias in the data, in the data system that it's trained on. And then you've got the whole issue of data integrity, you know, data integrity and data protection. Because once you put it into a ChatGPT or a Gemini or wherever, you're starting to, the models are going to learn from that. So, you know, having some safeguards in that, and again, taking it slow, Seeing it as a pilot and experimenting with it, I think is necessary. But I would not cannonball into the deep end of the pool yet. I'm going to get there first. That's just right.

JJ: Have you seen specific types of pilots people are trying or no?

EL: Yeah, I have. I've seen, it was interesting, I was working on some strategy creation processes. And one of the things that we were looking at is how, using AI as an interview agent across different people in the commercial team to get some voice of the market data, voice of the customer data, and to be able to synthesize that. So actually using AI to conduct the interviews and then synthesize the data where the AI can say, ask me 10 questions, 20 questions, you start them with a script and then say, do you have sufficient information now to help me formulate a competitive marketplace study? You know, the things that we would normally go mechanically do to say, I love like Mary Meko charts and, you know, looking at those Meko charts and looking at the growth rates in the industry, AI can take a lot of that, the mechanical piece of the interviews, and then actually use some of the data and synthesize it and analyze it faster than we can, and it's really good at spotting trends. It's interesting. It's a very, very minor use case, but I just did, I try to be a lifelong learner. And so I try to read a lot. So what I did is I took my book list of things when I started tracking them on Goodreads. And it just shows, here's the books, here's the articles, up here are the summaries I've written on the articles. And I put it in AI. And I said, AI, I want you to look at what I'm reading and have read over the last, and I've only used Goodreads for maybe a year and a half, two years. There's only, probably 150 books in there. But I said, what are my blind spots? What are the trends of what I tend to gravitate towards and read? And what am I missing that's going to shape my thinking? And then help me formulate an intentional plan for my own development so that I'm being a well-rounded thinker as opposed to reading the same things over and over and just recreating my own bias. And I can use AI as an impartial person to say, here are the data sets. It's things I can't see. So that's just a simple example for me. And again, that's more in the vein of lifelong learning, but it's a simple use case.

JJ: How about have you seen much of it yet on like the manufacturing floor? We want to improve this line.

EL: Yeah, that's a good question. So I see a lot of application in the daily management processes. You know, I'm a huge believer, you know, in breaking monthly goals and quarterly goals and yearly goals down. And, you know, we always use the expression, you want to make the, you know, make the month, you got to make the week. If you want to make the week, you got to make the day, make the hour, and that's why we do the work breakdown from there. At the daily problem-solving level, now I'll be honest, I'm still a believer in manually entering data and collecting data, and there's some research that supports why you should do that. But a lot of companies have gone down a digital analysis on lean daily management, or on the daily management data collection process. Where AI can help with that is it helps you start looking at the trends across the data. And it helps you see, is there something else that I'm missing here? Why does this keep coming up? Or this was an issue six months from ago or six months ago and it popped up again. And it can help you actually track and automate some of the accountability tasks. Now, I still like the manual post-it notes and check it day in, day out. I had some good There's something about.

JJ: Holding a pen or a piece of chalk or something or another.

EL: The research supports it, and it's the research actually comes out of the education arena, and it talks about retention of material that you read and study, and it says that you're better off taking handwritten notes. in order to get a connection between what you're engaging with. Because if you type the notes, what happens is you end up transcribing what's being said as opposed to synthesizing the key points. So the best practice, even if you want to type them, that's fine, but you should actually rewrite those notes. What does it mean for me? So that's why I still think there's a parallel. It may not be fully vetted yet, you know, to say, retention is better if a person colors in red versus green on a quality chart on a daily process. But I would tell you, I think it's directionally correct. And I can't tell you how meaningful it is when an operator on an assembly line can say, here's what the data is. I was red for quality this last, you know, in the last shift and here's why. as opposed to having a computer screen, show them the data, and they're like, it's just noise to me. I had a person years ago, Gary Cohn, taught me Six Sigma, and Gary was out of Motorola, and I remember him telling me, he asked the question, he goes, you know what the best companies in the world and the worst companies in the world have in common, Eric? And I said, no, what? He goes, they both use SPC. And the difference of what he was talking about is he said some people actually use it to make a difference. And some people will use the SPC charts to make, the Shuhart and Deming, wanted it to be done. Other people will do it because they say, oh, I've got to have this so it's out there. If you use daily management charts and you're just pencil whipping the data or doing something to an operator and they're not using it as value added, it's just waste. And I've also got pictures and, you know, I still come back to I have pictures of digital boards with no signal. And my analog boards don't have a no signal problem.

JJ: Right, right, right. So now I'm thinking as a, you know, the person in a manufacturing company or plant that I've got to lead this, a lead transformation. I don't know how you feel about lean training. Who needs it? Does everybody need it? Does only a few people need it? So if you could kind of tell me what you think about that. But when you throw AI into the mix, does that change the whole thinking about who needs training at all or what kind of training?

EL: That's a great question. I would argue with you or argue, that learning speed is really a competitive advantage. And how quickly can people learn and adapt and grow? I do not think AI eliminates training. Again, I've come at that respect for people and continuous improvement principles. I think it can accelerate training and it can accelerate learning. But I think the only way it does it, though, is if leadership has made it safe and they're not fearing it. And that's where I think the psychological safety comes in a lot, because there is a lot of fear of AI. So I think people need to have the capacity to learn. I always use the expression, and I say it very tongue-in-cheek, you know, the gap principles, generally accept your accounting principles, view people as a cost as a business, and they view physical assets, physical capital as assets. And I've always argued that it's hard when you view people as costs because the assets, the capital expenses, actually depreciate in value. And I would argue that people appreciate in value because they have the capacity to learn and they have the capacity to grow and establish best practices. And AI, that's why I said you've got to go develop people. You've got to talk to them about it. You've got to teach them the basic mentalities and the process opportunities and how do you serve your customers better. Everything we do with Lean is all about take care of your customers, serve them better than the competition, and then figure out how, what is impeding us from producing the good or service that the customer wants. And we're trying to streamline the time it takes to get that good or service to the customer. And if we can delight them and take care of our customer, they will reward us with business. But that is still a function of people in companies doing that good and service. He has an enabler. That's as simple as I can put it. And it's a thought partner. It can help us learn better. It can help us analyze data faster. But we still need people. We still need people to do the work. And I would still argue that we have people who are trapped in suboptimal processes. And it's our responsibility as leaders is to equip them, teach them, train them, and develop them to improve their own processes so they can serve their customers.

JJ: So are you saying they, we're talking, so it's lean training hasn't changed, doesn't it doesn't change under AI? They're still learning sort of the same things. Do they need to even learn how to use AI, I guess, right? Would that be another I think there is to lean training plus how to use AI? But how does it also speed learning?

EL: I think it's, AI can look at, aptitude and it can look at how quickly there's somebody getting something, it can move on. there's a lot of things in the educational world and they use these algorithmic learning curves and it can actually look at and test your reading and everything. And I'm using a textbook for an engineering economy class right now. And I can actually, find homework questions as students read a chapter. And what it does, the AI works within that system to say, okay, you know enough on this, we can go on to the next one. Or if they don't, it can continue to feed additional questions in to help cement the learning if there's gaps in certain key concepts that you highlighted. So there's an extension of that into workforce training. to help people say, what do you need to know and are you getting it? And whether it be, compliance training, that happens yearly or specific job training, I think AI has a place to help us understand, do we have mastery? But somebody still has to develop what that training needs to cover. You know, it still comes down to, you know, even some of the Toyota Kata methodology. in training the workers and the coaches and explaining what's the main key elements we need to do. I still would argue that the training within the industry is still one of the greatest training programs ever developed. And the principles of that can actually be reinforced with AI. Now, I do think that you need to have, there's something to be said about a person and a coach and, you know, Having that personal interaction, because you can pick up on communication and nuance. It's the reason why if you want to go do physical training, I think having a physical coach is better than having an AI person say, do 10 reps of this, 10 reps of that, because they can't see the form. Now, eventually, will they be able to? Yes. But, you know, there's a lot of nuance that we'll never lose sight of. At least, I don't think we will.

JJ: Okay. So you touched on this, but I want to ask you more specifically. Company wants to embark on a lean transformation and it wants to bring AI into that transformation. Where would you suggest it start? On a certain task, in a pilot area, somewhere else altogether?

EL: Yeah, I think you look piloted, number one. Come up with a specific use case. And what you'll find, if you come back to, Malcolm Gladwell wrote years ago, he talked about The Tipping Point, and then it's actually the updated version, Beyond The Tipping Point, is a fantastic book, and it's actually new material too. So it's worth reading both. But he talked about, you know, the ideas of kind of having those early adopters and finding that group that wants to do that. Find the people who have an interest in it, and to show interest in it, and then carve out to say, help me understand the use cases for this. And then communicate with it through, do brown bag lunches with them and have them come in and say, here's what we did. And maybe you start with market research. Maybe you start with voice of the customer or new product development or new service. Or maybe you start in the innovation processes. Or maybe you start in, you know, you're like, we've got to get better at problem solving. How can we do this? Get a group of people, do a Kaizen event. and then brainstorm the Kaizen event and think through a future state environment to say, how can we use these tools? when we do a 3P Kaizen event, to lay out a new facility or create a new process, you're already bringing in thinking of like biomimicry and different things to say, what are examples in nature that we can look at that would help us think through ways to do a specific manufacturing process? But maybe you do the same thing and you throw in the idea of, is there a technological solution we can embrace? Is there an AI application for this? And then test that in a Kazen event. So I'm still a believer in Kazen, get a group of people together, lock them in a room, you know, put them on Kazen vacation, dedicated with a charter, cross-functional team, a facilitator, and use that Kazen event to explore. If you don't have anything else you can go do, do a Kaizen event to say how can AI come through and be used in our process and do a 3P event with it. If you want to do 3P, call us. Call me. Let me know. That's a straightforward event to do that. But that's a very powerful thing to do, to allow people to experiment, allow people to learn, and allow people to grow.

JJ: Okay. We are running out of time, unfortunately. So before we wrap up, is there any kind of one big message or one key takeaway you want to make sure the audience walks away with?

EL: Yeah, I think you see a lot of this and you see AI is very much technology driven. But I would argue to you that the better we get at science and technology, the more important human leadership becomes. And I would say that, there's an evolution, and you'd think of, I mentioned Frederick Taylor. Frederick Taylor taught us to, study the work. Fast forward many, many years, and then, after Deming and, Crosby and Duran, and the focus on, driving out fear and those sort of things, we get to the Toyota production system and lean. And lean says, you got to take care of respect for people. I see AI is a next evolution in thinking. It is more an augmentation. So we still have the scientific view. We still have humans and the humanity of different things where we're looking at the human factors. AI becomes an augmentation. And what we're looking at is AI is going to help us to enable people to think better, to think faster, to maybe think broader and deeper. and help us understand our own biases and see patterns. So there's a lot of good things that can come out of it with this, with AI. It's real. It's not going anywhere. I mean, it's all there's either got to be with it or not. So I'm on the side of experiment with it. I think it is a thought partner, but I'm the driver. It's the co-pilot.

JJ: Okay, terrific. So before we wrap up, I want to again mention that Eric has written several articles on Lean and AI that are at theindustryweek.com website. So I encourage you to jump over there and take a look. They're great reading and they augment this presentation. And with that, Eric, Thank you very much for coming aboard. I appreciate you joining me today.

EL: Jill, thank you so much. I appreciate the opportunity to continue to have the conversation. And I'm on LinkedIn, so feel free to reach out there. But thank you for the opportunity to continue to have this talk. So have a great afternoon.

About the Podcast
Great Question: A Manufacturing Podcast offers news and information for the people who make, store and move things and those who manage and maintain the facilities where that work gets done. Manufacturers from chemical producers to automakers to machine shops can listen for critical insights into the technologies, economic conditions and best practices that can influence how to best run facilities to reach operational excellence.

Listen to another episode and subscribe on your favorite podcast app