Chris Couch plays a unique role, having served as the senior vice president and chief technology officer of the automotive supplier Cooper Standard, and the chief executive officer of Liveline Technologies. Both organizations use AI to make products that ordinary consumers may never think twice, such as brake fluid and polymer seals for car doors.
Christopher Couch has more than 21 years of global automotive manufacturing experience. He was the Senior Vice President and CTO of Cooper Standard, responsible for R&D, product development and engineering, product strategy and program management. He is also responsible for the profit and loss management of Applied Materials Science, a risk business unit dedicated to the commercialization of unique materials developed by the company. Couch also leads the CS Open Innovation Program, which aims to position Cooper Standard as the partner of choice for start-ups, universities and other vendors for open innovation.
Your comments are critical to the success of me, myself and AI. For a limited time, we will download MIT SMR’s best articles on artificial intelligence to listeners watching the show for free. Send a screenshot of your comment to smrfeedback@mit.edu to receive the download.
In episode 5 of season 2 of the “Me, Myself and AI” podcast, we discussed open innovation with Chris, automating rote memorization processes without replacing manual work, and attracting talent by fostering a startup culture.
For more information on how humans and machines can successfully collaborate, please read the 2020 Artificial Intelligence and Business Strategy report “Expanding the influence of AI through organizational learning.”
Sam Ransbotham: Things like brake fluid and chemical manufacturing don’t seem to be Gypsy-like artificial intelligence, but all of us may have benefited from AI and know nothing about it. Today, we are discussing with Chris Couch, senior vice president and chief technology officer of Cooper Standard, how we can indirectly benefit from artificial intelligence every day.
Welcome to “Me, Me and Artificial Intelligence”, this is a podcast about commercial artificial intelligence. In each episode, we will introduce you to people who use AI to innovate. I’m Sam Ransbotham, professor of information systems at Boston College. I am also the guest editor of the MIT Sloan Management Review (MIT Sloan Management Review) AI and business strategy big idea plan.
Shervin Khodabandeh: I am Shervin Khodabandeh, a senior partner of BCG, and I am the co-leader of BCG’s AI practice in North America. MIT SMR and BCG conducted five years of research together, interviewed hundreds of practitioners, and conducted surveys on thousands of companies to understand what is involved in building, deploying, and expanding AI capabilities and truly changing the way organizations operate. Required conditions.
Sam Ransbotham: Today we are talking with Chris Couch. Chris is the senior vice president and chief technology officer of Cooper Standard. Chris, thank you for taking the time to talk to us. welcome.
Sam Ransbotham: Why don’t we start by understanding your role in Cooper Standard. What are you doing right now?
Chris Couch: I am the chief technology officer of Cooper Standard. We are a global first-class automotive supplier. I am also the founder and CEO of an AI startup called Liveline Technologies, which originated from our research and development work within Cooper Standard. We provide components in the vehicle seal and housing space and fluid handling (whether it is brake fluid or coolant)-all fluid systems in the vehicle. We also invest in materials science and technology, which we believe will have an impact beyond the automotive sector.
The average consumer may not see many of our products. In fact, when we refuel around your vehicle, we hope that some of our products do not worry about them, but they are essential to the driving experience and having a safe and reliable vehicle. For example, we have developed a whole new class of polymers, which we call Fortrex. Fortrex provides better sealing around the doors and windows of the vehicle. Why is it so important? This is especially important when we enter the world of electric vehicles. With the reduction of engine and gearbox noise (because there is no longer a gasoline engine), other noise sources become more common, the biggest noise source is the noise caused by the wind around the doors and windows. By providing enhanced sealed packaging for these products, we believe that we have the right products for the electrified world.
Chris Couch: We spent a lot of time and money developing advanced polymer formulations. Historically speaking, many are trial and error. Industrial chemists often do this. We use AI to develop a system that can provide our chemists with suggestions on the next set of formulas to help them gradually try the final solution. In many cases, we have found that using this method can significantly reduce energy consumption. Dramatic means reducing these R&D cycles by 70% or 80%.
Sam Ransbotham: Very interesting. But before we talk further about Cooper Standard’s success in AI, can you tell us more about your own background and career path?
Chris Couch: First of all, I think the best way to describe yourself is to be a lifelong manufacturing addict. When I was a kid, I took apart everything in the house, and I might have been hit by wall currents more than once-I think I explained a lot to me today. I am the kid who built the first car with a tool kit. I am a hardcore mechanical engineer, focusing on school manufacturing and control. My auxiliary projects include the construction of drones that can fly at high altitudes. I’m just a manufacturing nerd. This is indeed the icing on the cake for my career.
I spent the first third of my working time in a company in Japan. I worked for Toyota and then joined Japan. I spent more than ten years designing and building factories with them, and finally participated in the operation of the factory. The second third of my career has spent time providing profit and loss for an automotive supplier, which is mainly from Asia, so I also have business dealings, which may affect a lot of what I said today. Then, the last third or so of my career was a CTO gig. I am the second one here. I am still an automotive supplier, but as long as we understand everything that is happening in the world today (whether it is materials science or artificial intelligence), we can get involved in various interesting technical fields. So here I am. If you asked me two years ago, I really didn’t expect to do my second job here, but it is really interesting and we are very happy to have some impact through these technologies.
Shervin Khodabandeh: Chris, please tell us about the open innovation of Cooper Standard. How is this going?
Chris Couch: You know, when I joined the company a few years ago, when we looked around our technology portfolio, first of all, I was overwhelmed by the areas that really needed to compete. I mentioned material science before, but there are different aspects of manufacturing technology and product design-the whole analysis and AI topics that we will discuss today-and firmly believe that there is no way we can do all the work by ourselves. Cooper Standard is not a small company, our revenue is only 3 billion US dollars, but we are not the largest company. Open innovation is actually an attempt to establish and attract ideas and technologies from the outside world, and even talents. In this way, we cooperate with universities and consortia around the world. We work closely with startups and use them as a source of ideas. In fact, if you wish, our first appropriate AI project is indeed achieved through open innovation channels.
We partnered with a new startup company called Uncountable located outside the Bay Area, and they helped us develop a system that can be effectively used as a consultant for chemists to create new formulations for the materials we have been using. If you want, this ultimately becomes a great accelerator for our R&D process, reducing the iterations in those design and testing cycles. That was one of those big “Aha!” moments-there is huge potential to improve yourself in many areas. We can’t do everything ourselves, so how do we really build external channels? We now call it CS Open Innovation, but this is the driving force.
Shervin Khodabandeh: It sounds like this is a very unique way to bring people with different backgrounds and different talents together and make them work together. Have you discovered the secret to achieving this goal?
Chris Couch (Chris Couch): I think whether it is artificial intelligence, whether it is material science, whether it is other fields, my answer is the same: it is actually related to the ability to focus. We (like many other companies) adopt innovation pipelines and processes that govern innovation, as well as stage checkpoint processes, because of focus. How do we quickly reduce the location where we will allocate valuable R&D funds, and how to manage this money correctly? Therefore, we think like a start-up company: we are making the least investment to answer the next most important question, or we will eliminate everything soon or put it into practice.
Shervin Khodabandeh: Quite a few fail quickly, test and learn, and make a big fuss about what is working, but make a big fuss about what is not working, right? Did I hear it right?
Chris Couch: Indeed, this is not the only one. I think there is nothing special about AI-based projects, right? We think in the same way and quickly try to inspire those who have a clear vision of ROI. Frankly speaking, I think one thing we have seen in analysis over the years is that the ROI of AI (especially when used in conjunction with manufacturing and Industry 4.0) is sometimes difficult to achieve. [There are a lot of ideas, interesting things related to data, but the question is, how does it translate to the bottom line? Moreover, even if this story cannot be told, even as a hypothesis that we will prove through innovative projects, it is difficult to justify the work.
Sam Ransbotham: But, it seems to be the opposite. If you pay too much attention to ROI, just a little later, where will you get something weird, big and unusual?
Sam Ransbotham: How do you maintain a balance between focusing on ROI and not missing opportunities [opportunities] or not pursuing incremental aspects?
Chris Couch: I think the mentality of the stage gate is very useful here. I think in the early stages, we will see a lot of crazy things. Through open innovation, we will generate crazy ideas. We have crazy ideas from our own team, which is great. We will look at them without hesitation, and even spend a little pocket money to catch up with them to a certain extent. So the question is, what will we invest in to try production? If you want, it’s really the next door.
Therefore, it is absolutely important that exploration is very important. Of course we can do this. What I am hesitant to say is, but it has some space to deal with ideas and technology, but when it comes to productization, you must keep a clear attitude and get benefits from it.
Sam Ransbotham: For AI methods, this seems to be different. I mean, you just said: “Well, AI is no different.” But… I think I want to know if there are some differences in these new technologies. These things may require more freedom than some methods. Can do some weird things other.
Chris Couch: I think this is fair. Based on our experience, I think one of the differences with AI is that you may not be familiar with these tools and applications for general technical personnel. If you are talking to a design engineer, or a manufacturing process engineer, they may have read something, maybe saw an interesting demo somewhere, but may not be proficient in how it works, let alone nuts and nuts Up. The cost of scaling at the enterprise level. Because getting the model from the CSV file on the hard drive to run in Jupyter Notebook is completely different from producing it on a global scale.
Chris Couch: I think a lack of understanding of technology makes it different. If we are talking about traditional robotics, or maybe a simpler type of IoT concept, then many engineers have good clues, maybe they have used something in their careers, but less in terms of AI . I agree, there is a difference. The good news is that I strongly believe that one of the beauty of AI is its low price. I’m just making a silly example of Jupyter Notebook and CSV files, but this is a great way to explore certain concepts, and besides gaining knowledge, the cost of doing so is almost zero. Even then, I think we have repeatedly proven in our internal team that even the price of knowledge acquisition is reasonable if you wish.
Shervin Khodabandeh: Chris, I think based on what you said, that AI is relatively cheap. I agree with this because of course we have seen the proliferation of proofs of concept, and different teams have tried different methods and different ideas. There seems to be a fact that AI is also difficult to scale.
Shervin Khodabandeh: I want your response, “Some things are easy to operate and difficult to scale. After scaling up, a truly meaningful return on investment will appear.” How do you make the transition? How do you make things easy to pilot and arouse excitement, but it is difficult to truly integrate into business processes and working methods? What do you think of this transitional work?
Chris Couch: Okay. Yes, this is a big problem, it is definitely not easy, maybe not because of the faint-hearted? Because sometimes it does require a leap in faith and expansion capabilities. The best I can say is our experience on Liveline: we did some very early prototyping, we thought we knew the data science side, but this was just the beginning. That was almost two years ago. Only in the past few months have we started to expand globally. The only insight I have is that when you prototype and experiment, you must choose practical use cases as wisely as possible, and everyone can go all out to connect the dots. The return on investment at the end of the day.
Sam Ransbotham: How do you attract people. …Once these solutions are in place, how are they adopted within the organization? How do you get people to work in teams that once had human partners but now have machine partners?
Chris Couch: Using Liveline…The basic concept of Liveline is to automatically create automation for complex manufacturing environments. We are using machine learning technology to design a control strategy that we deploy to the production line to control machine parameters in real time. We think this is very useful for attacking various processes that are too complicated or too expensive to be automated. Our early success lies in the continuous flow manufacturing process, chemical conversion, polymer extrusion, and we think it is suitable for oil, gas, wire and cable And other fields.
When we first entered the factory for on-site production trials, one of my concerns was that factory personnel might see it as a threat. We are automating; there are sometimes negative connotations in terms of the impact on people’s work, etc. However, I think there are a few things that really make us attractive, and the hospitality is very warm. In fact, the factory is now doing its best to promote it.
My attitude is to truly democratize the information and what happens with the tool. For example, we have spent considerable effort to ensure that operators in the factory environment have screens that can view data streams that were previously inaccessible in real time. Sometimes they are data streams that we create for machine learning. We provide them with visibility. If they want, we can let them see the decisions the system is making. We also enable them to turn it off-if they are not satisfied with the operation of the HAL 9000 on their production line, please click the big red button. In addition, we give them the ability to bias it. According to their experience, if they feel that the parts manufactured by the system are too small, so to speak too thin or too thick, then they can skew them a bit.
I think, at least in the factory environment, this kind of exposure and opening the black box is critical for people to buy and believe what is happening. One of the things we learned on Liveline is the enhanced feedback we get. We have received several influential and useful ideas from people. These ideas are actually nothing more than looking at data flows and observing system decisions. They asked us great questions and gave us great insights. [They] came up with some new types of data that we might want to label, and once they really start to have some intuition about what we are trying to do with data science, they may be useful. I think, if possible, at least in this case, the kind of democratization of the system and the openness and exposure of courage in its operation have always been one of the factors of success.
Shervin Khodabandeh: Sam, I think it covers a lot of what we talked about in the report, involving different ways of human-computer interaction. There is no black box. Humans are allowed to ignore or prejudice, but-Chris, I want to ask you, but before I have a chance to ask you, you have reached the point. This is the feedback loop. I think the next question I want to ask is, if skeptics become more friendly to AI, or build more trust between people and AI, how does the feedback loop work? Can you comment on this?
Chris Couch: Okay. I will provide you with an anecdote from a factory in the southern United States. In fact, this is the factory where we conducted the final pilot of Liveline before we decided to deploy Lively globally. We first make the production line run in a so-called automatic mode. My goodness, I think it was about the third quarter of last year. One of the pilots’ criteria is that we will run some A against B runs. The concept is very simple: “During these four hours, we will run the system in automatic mode. During these four hours, we will turn it off and you can run the device as usual. Then, in a few days and Over the course of a few weeks, we will aggregate statistics on scrap rates and quality, as well as unplanned stops, and then we will accurately quantify their value.”
A few weeks later, we came to the first review point. When I was sitting in the team, they pulled up their chairs a bit, looked at their shoes, and said, “Hey, we have a problem. We didn’t close the B data of the system. “I said, “Why is that?” They said, “Because they refuse to shut down the factory again once the factory is started. They don’t want to run the system without the connection because the impact is so significant for them and help them Operated the production line so much better that I don’t want to run it anymore.” This is very consistent with the type of reaction we have seen with other pilots in Canada and the Michigan Technology Center.
Chris Couch: Yes, this kind of feedback is very reassuring, but I still think that from the beginning, our philosophy was to be really open and show people what’s going on, let them look at the data, Become a participant in the problem. Solving, adjusting and enhancing have indeed laid the foundation for emotional connection and commitment to the project.
Sam Ransbotham: Those seem to be very different ways of getting feedback from the system, and then the other way you mentioned is to suggest re-introduction of new tags or new data.
Sam Ransbotham: For example, I can see that the deviation is adjusted to a kind of real-time feedback. Obviously, I want to press the red button to happen immediately. This is the meaning of the red button.
Sam Ransbotham: How do these non-instant feedback [loop] processes work? How do you deal with suggestions for new data and labels? Is there a process surrounding these?
Shervin Khodabandeh: By the way, I’m sorry to interrupt your answer. This is Sam’s. To a certain extent, my chemical engineering background also came into being.
Chris Couch: As you can imagine, at least for Cooper Standard, most of our production lines are chemical processing lines. We are using different types of compounds and are extruding them, if it is thermosetting plastics, put them in the oven stage-a process of 200 meters, and many of them are chemical processes. Yes, so you guys are in your best position.
Sam Ransbotham: What is the processing flow of the new data label? How do you standardize the process in a less real-time way?
Chris Couch: Let me give you a real example. About a year ago, we were doing an experiment. We had a process engineer. He was not a machine learning expert. He was observing the operation of the system, studying the data, studying the analysis generated by machine learning and how to perform the analysis. The predictable result is. At that stage, we did not get the desired result, what we saw was the change in output in the real world, and we did not get the prediction in the silicone world.
When he looked at the line, he said, “Look, I have a theory. I have a theory that one of the raw materials we are producing is changing. My theory is that this material is more susceptible to transportation to the factory. Experienced temperature and humidity history. When we ship nationwide, why not throw some data recording equipment on these palettes and be able to view the current and temperature history and integrate it into the analysis, see See if it can help us improve our predictive capabilities?”
Look, this is actually helpful. This is a real example of non-AI experts interacting with the system and using their human judgment to come up with improvements, even if they cannot write AI code. Once we make them sufficiently aware of what is happening, they can have some human intuition about what is happening here, and then they can participate in the process. This is a very powerful thing.
Shervin Khodabandeh: Chris, I want to ask you about talent. You have been talking about a lot of innovations, a lot of cool ideas-different groups inside and outside, come together to really try new things, try new things, and have a real game-changing impact. How do you think you can get the right talent, motivate them, keep them excited, and how to make the virtuous circle full of passion, vitality and innovation continue?
Chris Couch: This is a very good question. I think the answer may be different, depending on what kind of technical talent you are talking about. The way we think about manufacturing process engineers or control engineers may be a little different from the way we think about people with different skills in the AI world, sometimes in different parts of the country. I am not sure if there is a “one size fits all” answer. I think that, in general, when we find people who want to join the company, I think it will help a lot if we can show them that our ongoing commitment to innovation and doing cool things is real. I think being able to prove to people that you are willing to stick to the direction you invest in is part of the story.
Then, the second thing I think is important is culture. Make people believe that in addition to investing in resource availability, we also attach great importance to innovation. We just do good things seriously. From the conference room to the workshop, we are committed to winning through technology. If this culture is real, people will know, if it is not real and you are forging it, I think people will know. You cannot make money in a quarter; you must make money in a few to a few years. I think we have done a good job in this area, but this is really my key.
Sam Ransbotham: Chris, thank you very much for taking the time to talk to us today. You made some very interesting points. Thank you for taking the time.
Chris Couch: Very welcome. Hope you can tell me that I am very excited about AI. I am very happy what it can do in manufacturing and other industries. I think this will be an interesting future, and [am] look forward to [helping] build it.
Sam Ransbotham: Shervin, Chris made a lot of points. What makes you particularly noteworthy?
Shervin Khodabandeh: I think that is very, very insightful. Obviously, they have done a lot of work in AI, and made a lot of innovations and excellent ideas, and put many of them into production. I have been talking about the many key steps we take to get value from AI, which echoes what he said. The concept of experimentation, testing, and learning; the culture and importance of allowing people to try ideas and fail quickly and then move on; focus on some scalable concepts—focus on many things for testing and prototyping, but focus on some for expansion And investment things. I think that is really interesting.
Sam Ransbotham: I thought Chris was both excited and patient at the same time. I mean, [he] is obviously excited about some of the things they are doing, but at the same time, some initiatives took about two years to come true. It must be difficult to balance-get excited about something, and then wait two years for it to come out. I think this is a good integration.
Shervin Khodabandeh: Also, at this point, the importance of attention. I mean, once you choose it and make sure it’s the right approach, you will see that it is moving in this direction and realize that you shouldn’t give up now—you just need to mobilize and double up
One thing that impressed me was the importance of culture and how he said: “From the board of directors to middle managers, they must believe that we are behind, we are investing, and this is not just a fashion.” This must Infiltrate the entire organization to make the talent really excited and interested.
Sam Ransbotham: Even the people who use the system are involved. I think this is a good example. These people may have been so busy with their work that they can’t take a step back and think twice. He gave a good example of how to have the freedom of a machine to complete some work and let people do things that humans are good at. In our previous report, he covered almost all steps related to the different ways of machine operators. We did not prompt him to do so.
Shervin Khodabandeh: Another thing I really like is his view of talent. I asked him: “What does it take to recruit, train, and retain outstanding talents?” He said: “This is not a one-size-fits-all approach.” Recognizing that not all talents have the same coat, different people, different skill sets, and Different sensitivities, and they are looking for different things, but the common theme of the people who go there requires continuous focus and investment in innovation, and they want to see this, maybe this is commonplace. In this way, data scientists, technicians, chemists, and engineers may have different career paths and career ambitions, but they all share a common innovation effort.
Sam Ransbotham: I don’t think he mentioned him, but Chris is a mentor to Techstars. I’m sure that some of his background will also affect his views on different people and different ideas and how this talent is gathered.
Sam Ransbotham: Thank you for joining us today. Next time, we will talk to Quuiming Qu about how Home Depot will continue to build its AI capabilities. Please join us.
Allison Ryder: Thank you for listening to me, myself and AI. If you like the show, please take a moment to write us a review. If you send us a screenshot, we will send you MIT SMR’s best article on artificial intelligence for free within a limited time. Send a screenshot of your comment to smrfeedback@mit.edu.
Sam Ransbotham (@ransbotham) is a professor in the Department of Information Systems at the Carroll School of Management at Boston College and a guest editor of the MIT Sloan Management Review’s Artificial Intelligence and Business Strategy Big Ideas program. Shervin Khodabandeh is the senior partner and managing director of BCG and the co-head of BCG GAMMA (BCG’s AI practice) in North America. You can contact him at shervin@bcg.com.
“Me, Myself and AI” is a co-podcast of MIT Sloan Management Review and Boston Consulting Group, hosted by Sam Ransbotham and Shervin Khodabandeh. Our engineer is David Lishansky, and the coordinating producers are Allison Ryder and Sophie Rüdinger.
You must be logged in to leave a comment. Is this the first time? Sign up for a free account: comment on articles and get more articles.
Post time: May-20-2021