# How AI Changed a 150-Year-Old Industrial Manufacturing Company! | CXOTalk #901

- **Show:** CXOTalk
- **Published:** 24 Nov 2025
- **Source:** https://podcasters.spotify.com/pod/show/cxotalk/episodes/How-AI-Changed-a-150-Year-Old-Industrial-Manufacturing-Company---CXOTalk-901-e3bdc0p
- **Transcript:** https://podcasts.opennash.com/shows/cxotalk/how-ai-changed-a-150-year-old-industrial-manufacturing-company-cxotalk-901-c7ea8a/
- **Provider:** OpenNash Podcasts (machine-generated)

## In brief

Hexion CEO Michael Lefenfeld argues that a 150-year-old industrial manufacturer can become AI-first by rewarding people for using AI rather than pushing it on them. AI is a tool to lift every employee, speed up research, and help customers run plants with less waste, all with humans checking the results and safety coming first.

The central tension is that Hexion's AI helps customers buy less of its products. Lefenfeld argues the data gathered in the process lets Hexion tailor its products and services to each customer, deepening the relationship and avoiding the commodity trap. A second tension is moving fast on AI while never compromising on safety.

## Key insights

1. **Hexion rewards AI adoption instead of forcing it** The company ran an AI challenge with a $50,000 first prize and says there is 'very little stick' and 'lots of carrots'. Employees who resist can move into subject-matter expert roles that help train AI systems.
2. **AI recommendations beat full autonomy in plants** Hexion's smart tech software suggests process changes to operators, who keep control. Lefenfeld says it also trains operators, much as a glucose monitor taught him how food and exercise affect his blood sugar.
3. **Helping customers use less can deepen loyalty** AI tools cut customers' use of wood, energy and resins, but the data collected lets Hexion design formulations for each customer's process. Lefenfeld says this builds a closer relationship than competitors can offer.
4. **Private equity ownership enabled long-term AI investment** Lefenfeld says owner American Securities encourages a 100-year view rather than focusing only on its hold period. It also lends Hexion a roughly 100-person operations group, which lets Hexion experiment without hiring.
5. **Speed matters, but safety is never negotiable** Lefenfeld treats most decisions as reversible and says to try things quickly, but Hexion handles hazardous chemicals at high temperatures and pressures. Humans therefore always check AI outputs, and he says he will 'never be cutting edge' on safety.

## Questions this episode answers

### Why would a company use AI to help customers buy less of its product?

Lefenfeld says that reduction will happen anyway, whether Hexion or a competitor provides the tools. By doing it, Hexion collects data on each customer's process and can design better, customized products and services for them.

### Will AI at Hexion cost people their jobs?

Lefenfeld says Hexion is not looking to cut headcount and that he is not a 'grow by cutting' CEO. He says an employee who builds an AI tool that replaces their own task would more likely get a promotion.

### What is the 'golden batch' in manufacturing?

It is Lefenfeld's term for the perfect production run. AI helps get there by lowering raw material use, energy consumption, emissions and scrap while keeping quality high.

### Is Hexion's company data used to train public AI models like ChatGPT?

No. Lefenfeld says Hexion's AI systems are kept internally behind firewalls and are not publicly trainable by providers such as OpenAI or Gemini. He wants the systems to keep learning from Hexion's own data.

### What does Hexion actually make, and where is AI applied?

Hexion makes adhesives for wood panels like OSB and plywood, plus fertilizers, oil and gas products, coatings and paints. AI is used in research, in plant software for wood panel makers, in the supply chain, in reporting and in customer service.

## Facts

- **Michael Lefenfeld** [0:00:22]: "We make the adhesives that go into wood panels, so you think OSB plywood, it's a bunch of pieces of wood glued together."
- **Michael Lefenfeld** [0:01:41]: "Hexion's 150 year old company. We've got a lot of materials, a lot of processes, a lot of formulas, a lot of, you know, excellence in what we do."
- **Michael Lefenfeld** [0:07:05]: "We did an AI challenge at Hexion. First prize was 50,000 to a person or to a team that came up with an AI solution to one of the problems that we have."
- **Michael Lefenfeld** [0:12:35]: "We're not looking to cut headcount, that is not in our strategy at all."
- **Michael Lefenfeld** [0:13:31]: "I was chosen by the board to come into Hexion through a transformation for growth. My background is entrepreneurial, my background is growing new products, I have patents"
- **Michael Krigsman** [0:14:42]: "And I will mention you have over a hundred patents and pending to your name as well."
- **Michael Lefenfeld** [0:17:38]: "We have all human interactions still, you know, to double-check the results and make sure that we're not leading anything in the wrong direction."
- **Michael Lefenfeld** [0:21:36]: "AI enables us to run towards what we call the golden batch, the perfect batch."
- **Michael Lefenfeld** [0:23:09]: "We have parts that right now, we have we have overlays that go into the press, which is the heat and pressure that gets applied."
- **Michael Lefenfeld** [0:23:41]: "And then we have a product called the Smart Strander."
- **Michael Lefenfeld** [0:24:44]: "Instead of putting it into a data historian where it's gone all these years. It now brings analysis, interpretation, and suggestion for it to run properly."
- **Michael Lefenfeld** [0:26:33]: "Some operators are brilliant when it comes to that gut feel interpretation, as they're seeing different trends along their production line."
- **Michael Lefenfeld** [0:26:33]: "I'm a type one diabetic. Um, I've been a diabetic for, you know, almost forty years, and let's call it about a decade ago."
- **Michael Lefenfeld** [0:30:43]: "We have enabled our teams with tools to be able to reach through to our SDSs, our safety data sheets"
- **Michael Lefenfeld** [0:32:23]: "We're not public, we're a privately held organization, but, you know, public organizations as well are constantly churning that information for the needed shareholder."
- **Michael Lefenfeld** [0:33:17]: "They are kept internally, they are firewalled, um, and we do make sure that our data and our know-how and our, you know, and our intellectual property stay very secure."
- **Michael Lefenfeld** [0:47:04]: "American Securities is our PE owner, and they are absolutely fantastic."
- **Michael Lefenfeld** [0:47:41]: "they have an entire operations group, a hundred people that they provide to us, they pay for that allows us to borrow from those people."
- **Michael Lefenfeld** [0:48:30]: "They own AI companies now, going and buying AI companies."
- **Michael Lefenfeld** [0:50:18]: "The adhesives in the wood panel industry, in the wood industry, are made in what's known as batch chemistry. So big stir tank reactors"
- **Michael Lefenfeld** [0:51:54]: "The business model has not changed just yet, but we do see there will be ways to expand the business model between, you know, uh, with a service component"
- **Michael Lefenfeld** [0:53:44]: "Every great idea is one in a million, but there's 8 billion people on the planet, so at any moment in time, 8,000 people have that same exact great idea."

## Transcript

**[0:00:00] Michael Krigsman:** Why spend millions on AI that helps customers buy less of your product? Sounds like bad math, but Hexion CEO Michael Lefenfeld says it's the way to escape the commodity trap. I'm Michael Krigsman. Welcome to CXOTalk number 901. Let's get into it.

**[0:00:22] Michael Lefenfeld:** Hexion is a company that every person pretty much in the world interacts with every day, but they just don't know it. We make products that make products better, as BASF used to say. We, uh, we make the adhesives that go into wood panels, so you think OSB plywood, it's a bunch of pieces of wood glued together. We make all the adhesives around the world among, you know, among that space. We make fertilizers for agriculture, we make oil and gas, you know, purifying products.

**[0:00:57] Michael Lefenfeld:** We make coatings and paints and things along those lines that make those products better. So we touch the surfaces, the materials, everything you interact with every day. Um, has some hexion, you know, brand to it.

**[0:01:13] Michael Krigsman:** So you're kind of hidden behind the scenes, as it were.

**[0:01:16] Michael Lefenfeld:** We are. Hexion is a B2B business, so Hexion doesn't end up on a product name. You know, when you're buying it, but it is enabling the product you're using.

**[0:01:27] Michael Krigsman:** What I found so fascinating is your emphasis on AI. Give us some insight into what is that, why are you using AI, how are you using AI, and so forth.

**[0:01:41] Michael Lefenfeld:** Hexion's 150 year old company. We've got a lot of materials, a lot of processes, a lot of formulas, a lot of, you know, excellence in what we do. And AI is an enabler to that. AI allows us to take the best and make it even better. You know, that is really the critical push for AI, and we use it to look at all people in our organization. And, you know, as you can imagine, in any role, in any job, there's a bell curve of quality of people and performance.

**[0:02:21] Michael Lefenfeld:** There's some lesser performers, some really substantially productive performers, and then there's an average in the middle of that bell curve. Um, ranging from from good to bad. And what AI enables is taking everybody on the lower side of that bell curve and making them the best they can be if they use it in the right ways. And that's how we enable AI throughout the entire organization of Hexion, and how we also use it to make our products and the solutions for our partners and our customers even better.

**[0:02:56] Michael Krigsman:** Now, let's take a moment to learn about Emeritus, which is making CXOTalk possible. If you're a business leader navigating change or driving growth, explore executive education programs with Emeritus, a global leader. They offer programs developed in collaboration with top universities, designed for decision makers like you. There's a program tailored to your goals, whether that's AI and digital transformation or strategy and leadership. Find your program at www.emeritus.org. So it sounds like you are infusing AI through your operations in a in a both a a broad and a deep way.

**[0:03:46] Michael Lefenfeld:** Yes, every person at Hexion is empowered to utilize AI. You know, I used to when I first started here, I used to get emails from my associates in my organization and say, I did this work. I didn't use AI to do this, this was my work. And I'd write back to them and say, Why do you feel that that's a exception using AI and not the rule, right? When we go to the doctor and we ask the doctor to diagnose us, we want them to use all the tools at their disposal.

**[0:04:21] Michael Lefenfeld:** You don't want them just to, you know, just to listen to your heart and listen to your lungs and then guess what may be wrong with you, right? You want them to use every, you know, system and and equipment to make, you know, your diagnosis better. And at Hexion, we're the same way. We wanna use all the tools that are enabling for us to do the best work that we can. So AI needs to be championed at every level throughout the organization.

**[0:04:49] Michael Krigsman:** Can you drill down and give us some examples, make this concrete? And and the reason I'm asking it this way is because everybody talks about AI, there's so much hype and everybody says, Yeah, you know, we're investing in AI. And the words do not necessarily indicate the reality underneath. And I know for you, this is very serious, so please dig into this for us.

**[0:05:15] Michael Lefenfeld:** Innovation at Hexion is a critical growth lever that we utilize, and so if you think about an entire research organization, AI is an incredible enabler. There is a lot of new science published every day. There's hundreds, thousands of patents and papers and presentations and journals that that are given, you know, and and released on a daily basis. If I had my scientific team having to read every one of those papers, every one of those patents, every one of those publications to basically determine if it could be helpful through our innovation pathway.

**[0:05:59] Michael Lefenfeld:** They would never have time to be in the lab or in the, you know, in the research facilities, doing the work that we want them to do to make products better. And so we utilize AI directly in that role to scan a wealth of data that would never be able to be done by single individuals. And then summarize it. And then they can take that summary and say, Oh, this really hits home for what we're trying to do.

**[0:06:28] Michael Lefenfeld:** And they can then take that paper and read it more in depth, look at the figures, look at everything that's enabling it, it, and then allow them to take their strategy forward using that new learning. AI isn't a learning tool, and we want people to use it to educate themselves, think bigger, ideate farther, and really, you know, be capable of expanding their own boundaries.

**[0:06:55] Michael Krigsman:** Do you ever receive resistance from folks who are you're asking to, uh, adopt these tools and these new ways of working?

**[0:07:05] Michael Lefenfeld:** Of course, you receive resistance for pretty much everything that you try and put into an organization. That's change. Um, you know, change, manag- management is critical, um, to an organization and making sure you do it properly. And, you know, one way we got around that, you know, sort of reticence to adopt AI. We did an AI challenge at Hexion. First prize was 50,000 to a person or to a team that came up with an AI solution to one of the problems that we have.

**[0:07:38] Michael Lefenfeld:** And so we look at enabling people, whether they're subject matter experts, whether they're doers and coders, and, you know, data and analytics experts. We want all of them to participate in that AI push w- in the way that their strengths lend to it, right? In the people that have been in the job for a long time, you know, they've got a way of doing it, and they have a process that works for them. But it doesn't necessarily remain enabled, um, through the best technology that's out there.

**[0:08:14] Michael Lefenfeld:** Automation and manufacturing, you know, speed of data processing and being able to tweak the reactors and the and the products in a way that is using all the information. Instead of going into a report, but going into something that's giving you informed information to help you make decisions. And that's how you get them to really lean into it because they'll see the enablement that it causes.

**[0:08:39] Michael Krigsman:** Now, let's quickly hear from Emeritus, which is making CXOTalk possible. If you're a business leader navigating change or driving growth, explore executive education programs with Emeritus, a global leader. They offer programs developed in collaboration with top universities, designed for decision makers like you. There's a program tailored to your goals, whether that's AI and digital transformation or strategy and leadership. Find your program at www.emeritus.org. So there's a cultural shift that takes place. I I'm not trying to put words in your mouth where you are driving the use of that of the information that comes from the AI.

**[0:09:29] Michael Krigsman:** To actually take action steps and incorporate those learnings directly into the workflow.

**[0:09:37] Michael Lefenfeld:** That's exactly right. So AI doesn't have to be the doer of everything. Everything doesn't have to be made to be autonomous and automatic. What it can do is also a recommendation engine. It could say, Hey, this data is looking in a certain way. You may want to tweak a process condition to make something better. And this, you know, this is what also enables our smart tech technology that we commercialize out into the wood panel manufacturing industry.

**[0:10:05] Michael Lefenfeld:** We give users and operators that use our software recommendations to make their line work faster, work smarter, you know, use less material. That's where AI is really an enabler to complement it, you know, with the know-how of the people that have been doing the job for so long and have so much know-how and information. And pair it with the data processing power that is unmatched by a computer. So you bring those two together and you have a really perfect storm.

**[0:10:39] Michael Krigsman:** Folks, you can ask questions. If you're watching on LinkedIn, pop your questions into the LinkedIn chat. If you're watching on Twitter X, use the hashtag CXOTalk. We're speaking with the CEO of a large chemical manufacturer, Hexion. When else will you have the chance to ask someone like Michael Lefenfeld pretty much whatever you want? So I urge you, take advantage of it. And we have a a very interesting question from Twitter from Arsalon Khan, who's a regular listener.

**[0:11:16] Michael Krigsman:** Arsalon always asks great questions, and he says, What sort of disincentives and incentives do you have for people to adopt AI as a partner rather than a job taker?

**[0:11:29] Michael Lefenfeld:** We don't look at AI as a job taker, we look at AI as a, you know, is basically an improvement to the skill set. But we want every employee at Hexion and anybody that we hire to have some AI experience. And I don't mean they have to know how to build an agent, I don't know mean that they have to know how to code a language model. What I'm saying is is that they have to know how to, you know, utilize a proper prompt to get the information and be able to get their ideas and expense, you know, out there.

**[0:12:03] Michael Lefenfeld:** So that is a critical foundational lever for us. Now, if they're truly, you know, against using AI, you know, we can and and have found roles for them in subject matter expertise. To inform those that are more capable in the AI to build their systems, knowing what they do every day and how well they do it, and being able to translate it into that that that space. Now, we don't also at at Hexion, don't look at AI as a job replacer.

**[0:12:35] Michael Lefenfeld:** We're not looking to cut headcount, that is not in our strategy at all. What we're looking if somebody develops an AI, you know, agent or application that replaces their job, that would be phenomenal. Because that person will be an enabler for more parts of the organization to bring an AI thought process all the way through. So that person has probably a promotion on their hands, not a job loss, if they're able to do things like that.

**[0:13:05] Michael Lefenfeld:** So at Hexion, we use all enablement. There's very little stick, there's only lots of carrots when it comes to AI at Hexion.

**[0:13:14] Michael Krigsman:** Michael, you must acknowledge that there is real concern out there that as AI drives efficiency, there will be a temptation to eliminate certain jobs because you can.

**[0:13:31] Michael Lefenfeld:** I am not a grow by cutting CEO. I was chosen by the board to come into Hexion through a transformation for growth. My background is entrepreneurial, my background is growing new products, I have patents, I've, you know, I've brought new things to the world, um, throughout my career. So I am not the one that would be here to ultimately go through headcount reduction processes and and look at, you know, now, look, I will tell you, AI has enabled us to save money and do things better and more efficiently.

**[0:14:10] Michael Lefenfeld:** But, you know, the headcount at Hexion is not at risk from the AI, um, component. If anything, if we get, if we're freeing up time of individuals to do more, they'll do more. We have a lot of goals and a lot of plans in how we're looking to expand Hexion into, you know, doubling our size and and our and our revenue and our and our ultimately profit. But that is the goal for Hexion is just for growth, and that is what we are enabled to do, and that's the strategy we have.

**[0:14:42] Michael Krigsman:** And I will mention you have over a hundred patents and pending to your name as well.

**[0:14:49] Michael Lefenfeld:** I do, yes, I am a I. I am a nerd. It is the, uh, the single descriptor word that works, uh, almost perfectly every time. If you wanna describe me

**[0:14:59] Michael Krigsman:** It sounds like you have a different attitude towards AI adoption than many other CEOs who are looking at it through that lens of efficiency and reducing headcount as a result of automation.

**[0:15:16] Michael Lefenfeld:** through my entrepreneurial, early parts of my career, I've done every job. Um, I'm not gonna say I can do every job, but I had to do every job at the time. I wasn't great at all of them. But in any one day, I could be scrubbing toilets and meeting with a president or a CEO in the same day. So I spent time in supply chain, I spent time in procurement, I spent time in operations, I spent time in finance, all the different roles in the organization.

**[0:15:44] Michael Lefenfeld:** So, you know, as a CEO of, you know, now Hexion, a global, large, you know, billion dollar organization, I'm able to enable a lot of different functions with some very simple questions. Because I have an experience level in every role that I can be a little more intriguing, um, and lead to a little more creative thinking that comes from my background and my experience. And that's really what's been, you know, I think, what's taken Hexion to that next level as quickly as we have in the last couple of years.

**[0:16:24] Michael Lefenfeld:** You know, wood panel engineering is our and building and construction products are our, you know, are our bread and butter. We are leaning into that, um, you know, in a in a very big way, in a new way that hasn't been done before. And we think our partners and our customers are seeing that too.

**[0:16:44] Michael Krigsman:** We have a question for an interesting one from Nnamdi N. Agarkar, who says, What guardrails do you have? So when it comes to your AI, what are the how do you think about the guardrails?

**[0:17:00] Michael Lefenfeld:** AI does need some human interaction, at at least at this stage in time. Um, we don't let anything run without some, you know, some user feedback to it, um, but there are ways to build AI, you know, agents and models that do have its own checks and balances. But there is, I still think, you know, at this point in time, there is still a need for the human check. Um, and so we don't run anything at Hexion singly, singularly, um, with AI, you know, you know, exclusively.

**[0:17:38] Michael Lefenfeld:** We have all human interactions still, you know, to double-check the results and make sure that we're not leading anything in the wrong direction. Mind you, we we handle some pretty hazardous chemicals, um, as our starting materials. And, you know, we wanna make sure that we're doing everything right. We're working at with temperatures and pressures and different things to make, you know, to make different molecules. And because of that, we'd like to make sure that we have our, uh, you know, our human interaction in there as well.

**[0:18:10] Michael Krigsman:** So for you, it's not just the not just issues around AI ethics, but it's actually issues of, uh, relating to scientific accuracy and safety.

**[0:18:25] Michael Lefenfeld:** It's everything, right? You have to be thoughtful around any new technology. So I'm a I'm a very big believer in utilizing, you know, new technologies quickly and often. I also believe that a lot of decisions in, you know, a work environment, a lot of them are reversible decisions, and my team has heard me say this a lot. Um, if the if it is a reversible decision, um, so usually, if you know the co- of the the cost of the reversibility, it's usually time or it's money.

**[0:19:00] Michael Lefenfeld:** If you can afford that cost of reversibility, you should just walk through the door. Um, and if it doesn't work, you'll learn something, you'll be able to put it back together, and you'll build Humpty Dumpty back together again with a more learning and in a better way because of that learning. And that, to me, is a, you know, is a really important feature to being able to grow an organization with new technologies. That also being said, I will never be cutting edge when it comes to safety.

**[0:19:31] Michael Lefenfeld:** Safety is number one at Hexion, and safety should be paramount in every organization. So, especially with what we do, and even in AI, you know, capabilities, we wanna make sure we are safe in everything we do every day.

**[0:19:48] Michael Krigsman:** We have another question from LinkedIn, another interesting question. And folks, keep your questions coming. Take advantage of this opportunity to ask the CEO of a major manufacturing organization pretty much whatever you want. We're talking with Michael Lefenfeld, the CEO of Hexion, and this is from Anup Rama, and Anup says, How do you measure your ROI, especially when you're implementing AI in a chemical manufacturing environment?

**[0:20:25] Michael Lefenfeld:** There's numerous KPIs that come with that, depending on where you're using a, you know, using the AI, right? In in my team, in my back office team, it's productivity, right? They're getting so much more done in less time, they're able to do more, they're able to think more. You know, there's a I have a leadership coach that is that often tells me there's the. You have to spend time on the balcony as well as on the dance floor, and what that means is that you have to be able to see the forest for the trees at times.

**[0:20:59] Michael Lefenfeld:** And, you know, in my leadership team, the higher the leader in the organization, I want them more time spent on the balcony and less time on the dance floor doing the tactical work. Um, and then as you move throughout the organization to, you know, to a, you know, to the roles that are more tactical, they're spending more time on the dance floor and less time in the balcony. So AI really enables that capability, um, to be able to stretch yourselves into a, you know, into the the breadth that comes with an organizational, you know, work.

**[0:21:36] Michael Lefenfeld:** But if you're talking about AI and automation, you know, in in manufacturing, there's, you know, you see lower raw material consumption, you see faster batch times, you see better quality and better product. Um, at the end, you see lower energy consumption. It's truly sustainable, right? AI enables us to run towards what we call the golden batch, the perfect batch. And if you can lower the cost of raw materials, that's great for the for the customer. But if you can lower the u- the consumption of energy, or lower emissions, or, you know, decrease scrap or waste, that's better for the world.

**[0:22:23] Michael Lefenfeld:** And that's really what we care about at Hexion.

**[0:22:25] Michael Krigsman:** Let's drill into this point. How does AI help you lower these costs and drive towards that perfect batch?

**[0:22:38] Michael Lefenfeld:** I'll utilize our smart tech technology that goes to a customer, cause that, I think, is, you know, you know, a a great example for us. So we have a AI technology that gets ins that gets an overlay that gets installed into every unit operation. So the way that that a wood panel is made is you take a, you know, you, you take a tree, you debark it, you eventually, you know, you treat it. You cut it into small pieces or strips, then you dry it.

**[0:23:09] Michael Lefenfeld:** You add our adhesive to it, our resins to it, and then it goes on a mat, and it gets pressed with heat and pressure into the board that you see at the end, and then it gets inspected for quality. So our smart tech systems go into the specific unit operations. We have parts that right now, we have we have overlays that go into the press, which is the heat and pressure that gets applied. Now, that heat is generated by steam, so you wanna make sure you're using the perfect amount of steam.

**[0:23:41] Michael Lefenfeld:** So you're not generating any CO2 when you're generating that electricity or that heat, uh, that comes along with it. We have a smart quality product that basically makes sure that the product is in the, you know, is the right first time. So that you can go back into your line if product is not coming out well, you're not generating more waste, you're able to go back and fix in the plant itself. And then we have a product called the Smart Strander.

**[0:24:08] Michael Lefenfeld:** What a strander is is basically the knives that chop up the wood into the pieces that you need. So our AI software for that unit operation allows you to get the best con- consumption of that tree you can. So you're using less trees, you're having less waste, and you're making sure that the product quality is exactly where it should. Because you're making sure the knives on the cutter, on the strander, is not. They're not dull, they're not vibrating, they're not causing bad cuts, so you're making sure that every cut is the right cut.

**[0:24:44] Michael Lefenfeld:** So if you think about first time right, is what AI can enable by taking that data, that information, instead of putting it into a data historian where it's gone all these years. It now brings analysis, interpretation, and suggestion for it to run properly. Now, you can always run autonomously, but we prefer to run with a recommendation engine. We we do enable the the autonomous, but we prefer the recommendation engine. So operators and their expertise of knowing their line and how their line operates, make sure it works.

**[0:25:20] Michael Lefenfeld:** And this does lead to us using less consumables, which is, you know, which is a savings for everybody, the world, and for the customer.

**[0:25:29] Michael Krigsman:** So, in this case, the savings and the efficiencies derive from real time or near real time data analysis and the ability to make adjustments along the way. As opposed to increasing the knowledge, the general knowledge of, uh, the people working at the at at Hexion.

**[0:25:56] Michael Lefenfeld:** That's not exactly right. Yes, it uses data to make sure that you're putting your parameters in the best place possible, but it's also providing learning to the operators at the control board, right? So where they're sitting and adjusting and tuning and doing all the things that need to get done. Um, the AI tool is telling them, Hey, have you thought about this? And that's a training system, so you have to think about the AI recommendation tool is also making your operators learning, um, as they're performing the task because they're getting data analysis.

**[0:26:33] Michael Lefenfeld:** Now, they have a gut feel, right? Some operators are brilliant when it comes to that gut feel interpretation, as they're seeing different trends along their production line. But the AI engine can also now tell them, Hey, you didn't notice this. So let me give an example. On a personal front for me, I'm a type one diabetic. Um, I've been a diabetic for, you know, almost forty years, and let's call it about a decade ago. I started using what's known as a CGM, a continuous glucose monitor.

**[0:27:06] Michael Lefenfeld:** It's it's in my body and it tells me my blood sugar in real time. I run my blood sugars like your body runs its blood sugars, right? But I learned things that I thought was doing right all these years. When I ate certain things, my blood sugar reacted differently, right? It's it's a learning tool that says, Oh, if I eat that, it's gonna change my blood sugar in a different way than I thought. It takes longer to digest, it takes longer to absorb.

**[0:27:35] Michael Lefenfeld:** Sometimes exercise raises my blood sugar. You know, if my parents see this video, I could tell you my mother used to take me when before I had soccer tournaments as a kid. I they used to take me for ice cream, um, as a surprise, cause I was getting all this exercise. So, but what it turns out is when you start exercising, your blood sugar actually goes up. So she was giving me ice cream, which was raising my blood sugar, and exercise, which was raising my blood sugar, wasn't probably the best combination.

**[0:28:02] Michael Lefenfeld:** But I only learned that with a CGM. And so it's that type of tool now put into wood panel processing, chemical manufacturing, that gives you that learning ability that's different than, you know. So it's a trainer, not just a doer, and that's really critical.

**[0:28:19] Michael Krigsman:** So it's the combination of, uh, data analysis combined with learning, essentially. We have a question from Drew Jackson on LinkedIn, and Drew says, What use cases relating to customer service and client facing roles do you see AI having the largest positive impact on?

**[0:28:46] Michael Lefenfeld:** AI in customer facing and customer service, um, is actually pretty important. Um, as you can imagine, developing products for the customer comes with a with a skill of voice, a customer you wanna get all that feedback from them first, um, but also it helps with strategy, right? If you you can use an AI agent to, uh, or just a simple, you know, AI, you know, system to help yourself to go through the sales process. What kind of questions could be asked?

**[0:29:22] Michael Lefenfeld:** You could be you could sharpen your ability to do the sales pitch, to tune your your, your s, your, your capability to make sure that you, you know, highlight things that may be important to them. Cause you can use an AI agent to then say, Hey, what's most important to this customer? What have they been putting out in their press releases, their their pub, their their shareholder documents? What are their mission-critical, you know, goals for the specific year?

**[0:29:52] Michael Lefenfeld:** And then you can craft your sales pitch, you know, directly, so you can use it as a, you know, as a, you know, give and take to build that strategy on how to deliver, you know, sales excellence, um, around that. But you can also use it for strategic marketing, you know, and and data gathering. You use it for, like I said, for voice, a customer, um, for customer service help. So you have at your fingertips, you know, you know, you can gather that information.

**[0:30:20] Michael Lefenfeld:** That Hexion has hundred plus years of information on some of our products that we can pull that, you know, that dated information in a much more, you know, realizable way. To make sure that everybody can answer questions in the most informed positions.

**[0:30:37] Michael Krigsman:** To what extent have you applied automation in areas, uh, such as your contact center?

**[0:30:43] Michael Lefenfeld:** We have enabled our teams with tools to be able to reach through to our SDSs, our safety data sheets, right into the, you know, into the specif- the specific product stewardship sheets that are needed. We wanna make sure that they have permit access. They're able to reach much further than they used to into the trove of information that Hexion has for every one of its products and every one of its, you know, its solution offerings to our customers to make sure that they have the best service that they can give our customers, you know, immediately.

**[0:31:20] Michael Krigsman:** We have a question from Alfonso Velasco, who says, AI in manufacturing can it be applied both on the front line, the shop floor, and in the back office, accounting, sales, customer service, logistics? What are the most exciting use cases Hexion has deployed in the latter?

**[0:31:40] Michael Lefenfeld:** In the back office, AI is a critical tool in our supply chain excellence and our, you know, and our procurement excellence. To make sure that we're looking at especially in today's day and age where trade lanes are changing quickly, um, with all the geopolitical, macroeconomic movers that are happening in today's day and age. So using AI to enable those those pathways, um, to make sure that we can serve our customers without incident, without delay, without any challenge is an area where AI and, you know, that enablement really does, you know, s, you know, return on on that investment pretty quickly.

**[0:32:23] Michael Lefenfeld:** Um, and then there are just simple things, right? Reporting financial information using AI allows us to much, much more quickly put together reports and analyses that are needed for a large organization. We're not public, we're a privately held organization, but, you know, public organizations as well are constantly churning that information for the needed shareholder. And, you know, we use it, we use tools like that, you know, every day.

**[0:32:53] Michael Krigsman:** And this is from Greg Walters, who says, Okay, speaking of learning, instead of the AI teaching the human, isn't the AI itself learning beyond the human? And I guess we could extrapolate that question to the whole issue of the AI taking in your data, and do you have concerns about that?

**[0:33:17] Michael Lefenfeld:** The AI systems that we use are not publicly trainable to, you know, to open AI or or Gemini, or, you know, the systems we we. They are kept internally, they are firewalled, um, and we do make sure that our data and our know-how and our, you know, and our intellectual property stay very secure. So I, I, I'm not super worried about that as an issue, but look, the AI should learn. You want it to learn, you want it to learn on the data of our processes and our and our, you know, different actions and throughout the organization and every function.

**[0:33:53] Michael Lefenfeld:** And make sure that we're utilizing it into so that it's continually learning new scenarios, new ways we do things, new things we're looking at. Because as we're a growing organization, we may try to enter into different markets and different industries as we go about. So we want a learned system, we just don't want a uncontrolled system, and we want a s secure system around that.

**[0:34:20] Michael Krigsman:** You want a system that learns, but that doesn't share outside the boundaries of your organization.

**[0:34:27] Michael Lefenfeld:** Yes, and then we make sure we we put those firewalls in place as needed.

**[0:34:31] Michael Krigsman:** We have some great questions on Twitter. Isn't, isn't the audience fantastic? I mean, the questions we get are just wonderful. So, again, this is from Arsalon Khan. He says, A company is not an island. How do you encourage your ecosystem of vendors and partners to become AI-ready, too? And I'll just add to that one. And what happens if your ecosystem of partners does not become AI-ready to that extent?

**[0:35:03] Michael Lefenfeld:** As an organization of Hexion, as we transform into basically a cutting-edge organization, delivering new products and new solutions for our for our providers. Um, it's sort of car, you know, rising tide raises all ships, right? We make sure that we're n- we communicate what we, what we do and how we do it, and we share this information. And AI is an area that everybody is inquisitive about, and everybody sees value in. Um, I, I truly believe that is not a question, um, through the or through organizations now.

**[0:35:41] Michael Lefenfeld:** Some move at different paces, of course, but, you know, it probably also, you know, the AI use at Hexion does allow us to, you know, to move quick. And I think the those that are providing us, they see the growth that we're making, they see the differentiation we're, we're leading, and they wanna make sure that they could stay along. So you do see them try and adopt, or they ask questions on how they can adopt, and we partner with them so that we can help them transform as well.

**[0:36:13] Michael Krigsman:** Sounds like you are very much a carrot rather than a stick kind of leader.

**[0:36:20] Michael Lefenfeld:** I would say that, I'm not sure all my employees would say that, but, um, look, there there's a different in this area. It, you know, when you're trying to get people to grow, you're trying to learn, you're trying to give them the skills to be successful. You know, I, it's, you know, Steve Jobs said, the, you know, you know what happens if you know what happens if you're, you know, if you train your people and they leave, but what happens if you don't train your people and they stay, right?

**[0:36:49] Michael Lefenfeld:** And that's paraphrasing, but that is, it's. We want the best people in our organization, we want to enable them, we wanna pay for their education. You know, we do that regularly. We want them to be the best they can be, and hopefully they stay. But if they go somewhere else, we'll have a pretty good reputation, um, for for for great people. And I think that's just as important because the brand of Hexion is bigger than just the products we make, it's also the people that we bring forward into the world.

**[0:37:19] Michael Krigsman:** Very interesting how you talk about the brand as being beyond just the product, but it's really the the broader reputation of the organization of the company.

**[0:37:34] Michael Lefenfeld:** Yes, 100%, you know, Hexion's been around for 150 years. We weren't always called Hexion, but we've been around for 150 years. And the culture and the institution and the soul is not something that, you know, I can change. I can only help make it, you know, bigger and more modern to where it is today. And my team is amazing, I have one of the best teams in the world and they they don't stop.

**[0:38:03] Michael Lefenfeld:** Um, you know, we have a we have a saying at Hexion, you know, we strive, you know, we strive for perfection, knowing we're never going to achieve it. Um, because we're gonna constantly raise the bar and constantly try and achieve the next, you know, the next milestone. Um, and my team overdelivers every time.

**[0:38:22] Michael Krigsman:** Your respect for the DNA and the history of the company really does come through.

**[0:38:31] Michael Lefenfeld:** Thank you, I hope it does.

**[0:38:33] Michael Krigsman:** We have another interesting question again on Twitter, and this is from Russ Bankson, who says he's interested in acquiring and growing small manufacturers. How can he use AI to understand their current processes and suggest innovations to improve their processes and new products?

**[0:38:57] Michael Lefenfeld:** AI isn't a tool that should make that analysis and the diligence of those organizations in a really efficient way. You don't need a big team to do that, um, but you can benchmark really well using AI. You can see where the markets are going and where they're growing and where some headwinds or tailwinds exist. Um, you know, there are big private equity firms around the world that utilize AI to help see how, you know, how they may have bought something and made a mistake and why they made that mistake, right?

**[0:39:33] Michael Lefenfeld:** Or how they made an offer for a company that they said no to, or they didn't make an offer to a company and they said no to that. And then they saw it later become something that they'd never had in their model, right? So there there is a constant growth in that space for using AI and the information it can generate, um, and enable those to make more informed investment decisions. But I, I, I can't really opine more on that.

**[0:40:04] Michael Krigsman:** We have a question again, still on Twitter from Elizabeth Shaw, who says, What does being an AI-first company mean for Hexion today? And and what does that look like?

**[0:40:16] Michael Lefenfeld:** We want everybody to have AI as a enablement tool for themselves, right? So that's what it means. It means that everybody that comes into Hexion has to get comfortable with some use of AI. Now, it could be a native application in a software that, you know, we already have installed. It could be, you know, using a just a a prompt engine to be able to better, you know, better build their data, you know, acquisition. Or it could be building, you know, actual agents and actual tools in the organization.

**[0:40:51] Michael Lefenfeld:** But there is no real limit, um, as long as AI is a willingness in your, you know, toolkit. We want to give my job as a CEO is to give all of my team the tools to be successful. That is really the number one job. M in my day is to make sure that every person on my team has the tools to win. And AI needs to be one of those tools that they can utilize to help them win.

**[0:41:24] Michael Krigsman:** When you repackage AI or use AI for your customers in order to significantly reduce their materials usage, that cuts into your bottom line, right? You're selling them less. How do you reconcile your profit-making goal with your let's sell the customers less than we did before goal?

**[0:41:52] Michael Lefenfeld:** Michael, you s it sounds like you sit on my board. Um, you know, it this it, that's exactly right. We are putting tools in place at our customers to make sure they're using the right amount of our products, right? But that's going to happen, whether it happens by Hexion doing it or some other organization doing it, it's gonna happen one way or another. At least I have the tools now that will enable us to get better.

**[0:42:25] Michael Lefenfeld:** Because the information that we're ga- we're gaining by letting our customers use less wood, less energy, less resins that we sell, less waxes that we sell. Um, the reason that they can use less is because it's the right amount in the right conditions to make the best product. But that data then comes back to us, you know, it's screened, it's doesn't combine with other customers, but that singular customer then allows me to understand their process better.

**[0:42:59] Michael Lefenfeld:** So I can deliver them a formulation that is exceptional compared to anybody else out in the world that may try and supply that same product. Because I know intimately how their process runs, what the conditions are, what their substrates look like, what the surfaces may be, what the speeds they're running at. I can use all that information to custom design my products so they have something every day that is unique to their platform and allows them to have the best product in the world in what they're making.

**[0:43:38] Michael Lefenfeld:** Nobody can say they can do that, and I'm not saying we can do that just today, but we're on our way there and we are starting to show that return already, even in the early phase of our AI transformation.

**[0:43:54] Michael Krigsman:** It sounds to me that the calculation you're making, and please correct me if I'm wrong, it sounds the calculation you're making is fundamentally a trade. Which, on the one hand, we will use data to let you reduce your cost, let you reduce, become more efficient, can buy less from us and do an even better job because it's the right thing. In exchange, we are gaining data about you, the customer, that lets us serve you better than we could before.

**[0:44:35] Michael Krigsman:** And by knowing your business and what you need that much more thoroughly, we're assuming that we'll be able to, uh, be a better supplier, a better partner to you. And in aggregate over time, you the relationship will deepen between our companies, and in fact, you will, we'll sell you more.

**[0:44:58] Michael Lefenfeld:** Exactly, that's 100% correct. It's not exactly a trade, but we are trading those things. So hopefully, so we can expand our offerings to each customer so that they can have more from Hexion, more support. Because of what we're doing with the systems that we supply, with the, you know, with the product solution that we s that we, that we, that we offer, that we can expand that into their other parts of their business as well.

**[0:45:26] Michael Lefenfeld:** And that's how we look at it.

**[0:45:28] Michael Krigsman:** So, from your perspective, it's a deepening of the relationship based on shared data, essentially, that benefits both your customer and Hexion.

**[0:45:43] Michael Lefenfeld:** Yes, everybody talks about how they have an intimate relationship with their customers and their partners, but this is a much deeper intimacy. Because you're able to help them to be successful in the best way possible, um, and by ultimately giving the most sustainable solution they can possibly have. Because they're using less, they're wasting less, and they're bringing that to the communities that they serve and the markets that they sell into, all that much betterment.

**[0:46:17] Michael Krigsman:** And when you use the term sustainable, in this case, it sounds like what, what it, what it actually means is a combination of, uh, cost effective, as well as higher quality or better fit to purpose.

**[0:46:29] Michael Lefenfeld:** Yes, that's exactly right.

**[0:46:31] Michael Krigsman:** Michael, you are a PE-owned, private equity-owned company. Has that relationship changed the way you think of investing in AI? In other words, either has it restrained the amount that you could invest in AI, or has it forced an acceleration of that investment because the time horizon of private equity is limited and you need to show results?

**[0:47:04] Michael Lefenfeld:** American Securities is our PE owner, and they are absolutely fantastic. They have empowered me to think not just within their hold period, but to think about the business for a hundred years forward. They are not limited to the time constraint of, like you mentioned, the the the hold period where they look for their investment in the cash flow. They want us to build a business that is going to be long-standing for a future owner to baby basically be able to grow and continue to grow, um, forward.

**[0:47:41] Michael Lefenfeld:** So American Securities enables me to be incredibly, um, rapid in my movements. So, a, and I'll give a little bit of history here for American Securities. Um, they have an entire operations group, a hundred people that they provide to us, they pay for that allows us to borrow from those people. They're experts in legal, and they're experts in IT, and they're experts in procurement and supply chain. You name it. But what that does is it allows me to experiment and try different things without having to truly invest in hiring somebody, taking them away from a job they may be in, and then find out that idea didn't work.

**[0:48:30] Michael Lefenfeld:** And I would never wanna pull somebody out of an organization and then that, that, and then tell them they have to go look for a job. That's the worst thing a CEO could ever have to say to somebody. And so what American Securities enables us through their resources group is to experiment, and their IT team is phenomenal. They own AI companies now, going and buying AI companies. The valuations may not necessarily lean themselves to a perfect, you know, synergy with with with PE, but they own AI, they know AI, they're using AI, so they're really enabling and really looking forward to the solutions that we're providing because they then share it with the rest of their organization as well.

**[0:49:14] Michael Lefenfeld:** So I think it's been a really, you know, synergistic relationship between Hexion and American Securities because they are pushing for growth every way they can.

**[0:49:25] Michael Krigsman:** So clearly, the AI-first strategy that you've adopted fits with their thinking, and it sounds like you're very much mutually aligned around that.

**[0:49:35] Michael Lefenfeld:** Yes, very much so.

**[0:49:37] Michael Krigsman:** And I really encourage people to who are listening to think about this, which is the notion of using the data to create a win-win situation with your customers, where you are helping them be more sustainable. In exchange, that data enables you to develop a deeper understanding of their business, which crea- it's it's really a one plus one equals three because you're you're increasing the size of the pie based on that data.

**[0:50:18] Michael Lefenfeld:** That's exactly what it is. It's by us having that data, it allows us to generate products that are, you know, that are specific, that are designed for them. You know, the the adhesives in the wood panel industry, in the wood industry, are made in what's known as batch chemistry. So big stir tank reactors, um, opposed to continuous flow, where, like petrochemicals and oil and gas are refined, things like that. So because we do things in batch, we have to do sometimes to a specific wood and a a wood panel plant, we can send six deliveries a day, have to run three batches a day.

**[0:51:00] Michael Lefenfeld:** So with their constant change in, you know, environmental conditions, in their wood, shrubs, substrate conditions, with their, you know, with their line conditions, um, with their, you know, their. The heat conductivity in their press may get worse and worse over time, may need to do some modifications. We know all that, so I can give them a resin that's going to perform at the optimal conditions, and those optimal conditions make sure that they have the best product every time.

**[0:51:32] Michael Lefenfeld:** And that's the goal. And that data is symbiotic between us and them to make sure that we're all getting to the best place we can every single time.

**[0:51:43] Michael Krigsman:** We have a question again from Elizabeth Shaw, who is asking about the business model. How has AI changed your business model?

**[0:51:54] Michael Lefenfeld:** The business model has not changed just yet, but we do see there will be ways to expand the business model between, you know, uh, with a service component as not just a, you know, chemical materials, um, sales process. So we do see enablement in in, you know, in being able to deliver, you know, value in other ways through service and through other exceptional models, um, throughout. But I do see Hexion, and we already have started expanding beyond just the, you know, we sell you a product and you use it to go forward.

**[0:52:30] Michael Lefenfeld:** We are now getting into a more integrated, um, value proposition with our customers.

**[0:52:36] Michael Krigsman:** If you could go back to the start of this AI transformation, what would you do differently?

**[0:52:43] Michael Lefenfeld:** I can't tell you exactly what I've stubbed my toe on, but we've stubbed our toe plenty of times. But like I said, they're reversible decisions, and you learned from those reversible decisions, so I was able to go back, reverse and do it again in a smarter way. So, you know, I don't think, you know it, because AI is, L- you know, we're sort of building the plane while flying it, right? Cause AI is evolving so quickly, it is amazing the speed and the capabilities that are coming along with AI.

**[0:53:16] Michael Lefenfeld:** Um, so we're just, you know, we're, we're trying as we go, and but we're not afraid to make mistakes. That's okay, you know, you learn from them and you try again, um, as long as you're making sure everybody is safe along the way.

**[0:53:29] Michael Krigsman:** So many companies are stuck in AI pilot purgatory. What advice do you have for other business leaders who want to get out of the pilots and into practical AI benefit?

**[0:53:44] Michael Lefenfeld:** Just try it. Speed is important. It it's you really do have to keep up with what's going on in the world. It's moving incredibly fast. Um, and if you can't do that, you're gonna be left behind because somebody is going to, you know, I is going to move faster than you. There's a, there's a famous saying, right? Every great idea is one in a million. Um, but people don't do the math past that step. Every great idea is one in a million, but there's 8 billion people on the planet, so at any moment in time, 8,000 people have that same exact great idea.

**[0:54:21] Michael Lefenfeld:** So if you're not moving fast, if you're not executing quickly, if you're not bringing the smartest people to the problem and working as a team because you get there faster as a group, You're gonna get left behind because that other person, one of those 8,000 people is gonna be faster than you and you're gonna miss out on your opportunity. So to me, that is critical.

**[0:54:42] Michael Krigsman:** So in other words, just get it done. Get it done.

**[0:54:46] Michael Lefenfeld:** Just try it and get it done.

**[0:54:48] Michael Krigsman:** And with that, we're out of time. A huge thank you to Hexion CEO Michael Lefenfeld. Michael, thank you so much for being with us. I'm very grateful to you.

**[0:55:02] Michael Lefenfeld:** Michael, thank you so much. This was wonderful. I really appreciate it. And anybody who has questions, if they have more questions, can reach out to me on LinkedIn or reach out to our Markham team and we'll get you answers.

**[0:55:14] Michael Krigsman:** That's an invitation. And folks, if you're listening, uh, definitely connect with Michael on LinkedIn. Connect to me on LinkedIn. We have great shows coming up. Do not forget to subscribe to the CXOTalk newsletter. Go do that now. So we can notify you and you can participate. Thanks for everybody who watched and asked such awesome questions. You guys are fantastic. See you again next time. Have a great day, everybody. Bye bye.