What are the challenges that you see and, and that you guys are trying to tackle?
Oh man. Let's see. The market's insatiable. The customer base wants more, wants it faster, wants it in different flavors, wants it in different ways packaged differently sustainably, of course. So that's really hard as well to try to get more from the existing team, sometimes even a smaller team because of economic issues. So the old adage of how do we do more with less is, is very, very present right now. So that's, that's a big challenge.
The, the other challenges are that this stuff is really hard. You know, chemical development, material science development is hard. It's, it's multidimensional. It has an unbelievable amount of variables.
A key component of the modern world economy, the chemical industry delivers products and innovations to enhance everyday life. It is also an industry in transformation where chemical executives and workers are delivering growth and industry changing advancements while responding to pressures from investors, regulators, and public opinion, discover how leading companies are approaching these challenges here on the chemical show.
Join Victoria Meyer, president of Progressio Global and host of the chemical show. As she speaks with executives across the industry and learns how they are leading their companies to grow, transform, and push industry boundaries on all frontiers. Here's your host, Victoria Meyer.
Hi, this is Victoria Meyer. Welcome back to The Chemical Show. Today I am speaking with Ned Weintraub, who is the Chief Revenue Officer at NobleAI. Noble's a pioneer in science-based AI solutions for chemicals and material companies. And Ned has been involved in digital innovation for industry throughout his career at companies including Seven Signal Verana and HP Cloud.
We're gonna be talking about business challenges with accelerating chemical development, innovation, how AI fits into that and a whole lot more. Ned, welcome to The Chemical Show.
Thank you for having me, Victoria.
So you've spent your career in technology and growth oriented firms. What got you started in this space and what ultimately brought you to NobleAI?
Yeah, it's a good question. For me, I've always tried to work for Mission-driven companies, and Noble is a mission-driven company. We believe that AI can solve the world's problems better than it can destroy the world. And so we believe that that. This is, this is a fantastic way of, of really trying to solve health issues, environmental issues all sorts of different ways that we can, you know, get over what we're trying to, to fight every day.
Yeah. Makes sense. And I, and I like that mission driven space because I think we sometimes lose track of that in business. And yet it all starts out with a, a bigger purpose.
You got it. That's exactly right. It helps us get up in the morning.
Absolutely. So the chemical industry it is definitely in a period of accelerated innovation, right? We're seeing this, in fact, you and I recently met up at ACI, One of their themes was innovation. So innovation is everywhere, and yet we're still challenged in commercializing these new ideas, new products, new innovations. So first of all, what trends do you really see driving innovation in the chemical space when you go out and talk with customers?
I would say one of the biggest ones, was sustainability. Sustainability has finally arrived and moved from sustainability washing to really putting budgets behind it. Again, a mission. Where companies can feel good about what they're doing and really driving sustainability through everything. And frankly, product development and, and optimization is a big one. That's the biggest one. But I would also say adapting and agility to regulatory issues.
That was the other big focus and we hear this really from every single one of our customers, which is the world is changing fast. How do we adapt? How do we become more agile? How do we become proactive instead of reactive to a lot of these things? Those are the two biggest ones that we see kinda driving a lot of the the innovation.
Yeah, so I think the drivers are strong and as you say, they are mission-based in many ways, and yet there are still some challenges. What are the challenges that you see and, and that you guys are trying to tackle?
Oh man. Let's see. The market's insatiable. The customer base wants more, wants it faster, wants it in different flavors, wants it in different ways packaged differently sustainably, of course. So that's really hard as well to try to get more from the existing team, sometimes even a smaller team because of economic issues. So the old adage of how do we do more with less is, is very, very present right now. So that's, that's a big challenge.
The, the other challenges are that this stuff is really hard. You know, chemical development, material science development is hard. It's, it's multidimensional. It has an unbelievable amount of variables. Everything from price to market pressures to supply chain risk, personnel risk. This stuff is really hard. So how can technology like science-based AI help in these situations, and that's really why we're here.
And it seems to me, I mean, I think part of it, you talk about the people challenges and just the fact that it's really hard. I know what we're seeing is the continual graying of the chemical industry. Although some of us like better living through chemistry, so we wash that gray away and then what have you, right.
Absolutely.
But I think what I'm hearing from people across the industry is a real concern that they're losing really experienced staff. Yeah, across the board, and it seems particularly in the product development and formulation development space because they're retiring, right? So they're just, they're, they're moving on to greener pastures or however we wanna say it. Um, and, you know, we, we just don't have the same knowledge base in the industry.
And it seems like, AI is one of the ways that we can really harness and leverage some of that existing knowledge, even perhaps when the people that developed it have moved on.
Yeah. Unbelievable. I thought there's two issues. Number one, you're absolutely right. Senior leaders are retiring, right? They've been in this industry for a very long time. But the other one, which is really interesting is that there's a talent war. And people are leaving. It's not the old days of sticking with your company and getting a gold watch at the end of 30 years. The world has changed. So how can companies de-risk? You know, the personnel flight, the brain drain as they call it.
So is there a way to really mimic or really institutionalize or codify that institutional knowledge and not be at risk when that person walks out the door. That's also a really interesting conversation that we're having with customers. So, you're a hundred percent right there.
Right. So I mean, you guys are at the, the front end of moving AI into product development and developing solutions to support the industry. What do you see as the role and what are the conversations that you're having about how people want to integrate AI in this space?
Yeah. The best thing that's happened to us is generative AI, right? You know, the chat GPTs and the Geminis and all of these others have really brought it to the forefront. Now remember, artificial intelligence has been around 30 years longer than the internet has been around. So just just to give people an understanding, this is not, it's 15 minutes of fame. It's been around for a long time.
It just now has, because of the internet, has been able to take all of this data, the oceans of data that they can mine. What's great for creating speeches and my kids' homework doesn't work for science necessarily because there's not a lot of data. There's not an ocean's amount of data because we're trying to create things. And so, you know, if it were that easy, everybody would do it.
We have to be able to help our customers drive new innovation optimize current products with not a lot of data or completely spread out data. For us, this notion of specialized AI, or science-based AI, especially in chemical and material science is really critical because we can do a lot with very little data. We can be very prescriptive in what problems we we're being asked to solve and really be able to accelerate based on that.
That's a big piece of, it's great that we have generative AI but you know, specifically focused in this world that you and I live in, it has to be specialized.
Right. Yeah. It's interesting. So you say that there's not a lot of data. I think people would assume we are awash with data. And certainly it's true in, when I think about chemical companies and just their overall business and business operations, we have a ton of data about customers, we have a ton of data ton of data about manufacturing we probably have a lot of it about product and product development.
Although I will say, I reflect back and think about my time in industry, when I worked really closely with our formulation guys to help get new products out into market and stuff, I was sometimes shocked by like how many data points, how few data points were actually on a curve or on a graph or what have you. So it's it's interesting twist on this 'cause I sometimes think.
We feel like there's just all this data and yet maybe it's not always the right data or in the right place at the right time.
All of it. Absolutely. All, all of the above. You're right, these companies have been around for. You know, some a hundred years and you would think that the data is A, there and B accessible. And very often it's not. And neither one or one of them. And it's, it's spread all over the world in different lab notebooks, physical and electronic, maybe electronic. It's very, very difficult.
And so, how do we help customers get started without having to rake the ocean full of ones and zeros to get started? That's really where building that science into the AI right from the get go alleviates a lot of that need for a lot of that data. Because we already know a lot of that information. We can build it in institutionally and really accelerate that development.
So what does that look like? So say more a little bit about that. 'cause I think it stills kinda, you know, 20,000 miles up in the, the atmosphere. You know, what does it mean to be science-based? Is it generative? this generative? And how does it work?
Yeah. So it's, it is generative in a way because we can generate new insights, but fundamentally, when we talk about specialized AI or science-based AI, we're building the fundamentals of physics and chemistry into the models that we build with our customers in partnership. I always like to say for the senior executives who aren't necessarily scientists, and I don't have a scientific background, you know, elephants don't fly. We all know elephants don't fly.
But with, with commercial AI, something you get off the shelf, you have to train it, that elephants don't fly and therefore. It takes time. And then a lot of the answers early on you get, well wait a second, this doesn't make any sense. Why are we even going down this path? So they give up.
Where by building all of that knowledge upfront into the models and then using different models, solving different problems that really accelerates the insights and that gets us to, even in the first rudimentary models that we build with our partners. There are aha moments. We've, we've solved some very fundamental problems for customers that have been struggling with, you know, maybe it's a PFAS chemical that they're trying to get out of one of their formulations.
We were able to, you know, give them insights within 30 days, something that they've been trying for years to solve, or at least, you know, the last two years, within 30 days, we gave them directionally approaches to head to. So by building that institutional and, and that scientific knowledge into those models early on, we really accelerate that. And then we can train those models as we continue.
And then customers use them once they're mature to drive insights to really do a lot of testing that they wouldn't otherwise be able to do on a bench, if you will.
yeah. Makes sense. Yeah. Yeah. 30 days seems fast. I know. In the world of chat, GPT, if I have to wait longer than 30 seconds for my answer, it seems like it's taking a long time. But, but to your point, this is I guess a much more rigorous approach as needed, right? I mean, it has to be rooted in the scientific principles, whether it be chemistry or physics or material science to, to make that happen.
right. We're, we're dealing with scientists, by the way, so, you know, they, they want facts. They
They want facts and, and they're probably a little bit risk averse. So let's, let's talk about those risks. So what, you know, I think what are the risks that you guys see when, or that you talk about and you work to alleviate with your clients when you think about using AI and product development?
Yeah. So honestly, the risk that we see our people are gripped by this notion of not having all the data in one place. And honestly it's a bit of the boogeyman that, you know, part of the industry who is about trying to collect all of your data in one place before you get started. That's their message. And the reality of it is, is that the risk of not getting started now. You're allowing your competitors to, to distance themselves from you, right?
To either gain the edge or to expand if, if you can't do that, you know, part of the reason why we do a lot of work with mid-size companies is because they can't throw money and bodies at this internally, and they have to do what they have to do in terms of closing the gap with those big companies. And we see that every day and they genuinely see AI as.
Both a panacea, so we have to kind of temper their enthusiasm, but also give them the true value, you know, vision of what it can really do for them. So the risks, getting back to your question, are just getting started. That's one. The other risk is the industry has spent a hundred years hugging their IP and not allowing it out into the world because genuinely that is the keys to their kingdom. How do they work with partners, feel comfortable about working with partners, but still have the.
Security literal and figurative, figuratively, to be able to really collaborate with people outside their four walls. And that's a, that's a perceived risk as well, right? The cloud industry had to go through this, right? It's all of our data needs to be here. And then people realize that AWS and Azure are probably even more secure than your own network itself. So, these are early days. Those are the, those are the challenges that we work through with our customers.
I can see that. And certainly the ip, the intellectual property and data privacy is, is probably the thing I hear the most. Um, in many ways it's maybe the most misunderstood in my opinion. And, and I've done some work around this some folks that. You know, there's this perception of, oh, if I, if I put it out there, it's there for the public domain, Right, It's like, well, no, no, no, there are firewalls.
And by the way, don't put your don't put your test data into chat GPT because chat GPT is open domain, right? So buyer beware. But there are other, I mean, heck, there's a version of chat that you can buy that's private and obviously when, if you are working with a company like Noble, there's firewalls and privacy protections to protect. All that data.
Yeah. I'll, I'll even go one further. So you're a hundred percent right. There's the risk on the generative AI side that you are putting all of your. Your information out on the internet. So our customers do have and are working through these policies for their employees. So that should be, that should be looked at where science-based AI folks like NobleAI work. Is within their customers domain.
So we have the ability to build these models and serve them to our customers within their private cloud. So that's a big differentiator for us because we feel like, yes, we're not gonna try to change the hearts and minds about people's ip, right? It's, we don't have enough time. To, to try to, to create a sea change there. So for us, we feel like we, we have the ability to work within that. You know, that framework, and that's been very successful.
The second thing is, the other challenge is in the AI world is this notion of, of who owns the models, right? Who owns this ip, right? Is it the, is it the AI company or is it the chemical company that's bringing that data? And historically, all of the last five to 10 years. The AI companies have said, oh, no, no. Those are our models. Those are our models. And so it's really set up to be this very confrontational you know, is it ours? Is it theirs, is it co-owned? How do we do that?
If we wanna publish it, I mean, it, it can become a nightmare. We've taken a different approach. We, we build customized models specific to our customers, and they own those models because for us, it's, it's important that they can build from there. It's theirs. I always use the, the analogy of. Steven Spielberg writes a screenplay. He wants a movie made. He goes and raises money and go, gets it made. He takes it to a filmmaker.
He takes it to, you know Skywalker Rancher and they own the way the movie gets made, the special effects and all of the different ways that it becomes a fabulous Steven Spielberg movie. Steven Spielberg owns that movie. The movie company doesn't own that movie. And so we, that's our approach. We feel as though, and our customers appreciate that because they don't have to focus on. Oh my God, are they gonna turn around and sell this to a competitor?
Which is obviously in business a very real situation. So that's, that's how this industry is starting to evolve. We feel like we're on the forefront of it.
Yeah, that ownership risk, that's great. The other thing. That I hear, and this is a widespread concern regarding all AI and all generative AI, is that we're using a limited data set and that we're just kind of creating this very narrow spiral based on limited data. And once it gets skewed, the truth gets skewed.
Yeah. The bias, the bias comes in. Yeah. That bias is always a, an omnipresent thought around building these models in AI, it's the, it is one of the biggest reasons why companies should be partnering with companies like NobleAI because. Institutional bias happens within the same four walls. It's the same people building these models, and they, and they own that box.
Now, AI does a great job of, especially science-based AI and specialized AI to broaden your horizons, right, your design space, but. Bringing in people with external experiences, a from either people within the industry or even better yet, people outside the industry. What are the people in oil and gas doing? Although it's related, upstream is very different. What, how, what are they, what are they doing in exploration? What are people doing for alternative energy?
Is there something there that we can bring to the packaging industry? What are people doing in, in Biosynthetics and, and what can we bring to that? So, you're right, that's if you are trying to do it internally and without multiple different ways of building these models with NobleAI. I think for our, our domain of different models, we've got over 45 different ways of building models that could all be combined. It's not just kind of repurposing the same model over and over.
'cause that generates bias.
Yeah. Yeah. And I suppose once a solution is identified, let's just say you, you brought up the PFAS example. Once the new alternative formulation that replaces PFAS is identified, there's still lab work that gets done. There's all kinds of testing. And so new data is created
absolutely.
into the model. As long as I guess it, you know, you understand where it goes in and to your point, there's, there's always biases always existed in,
Always existed even more so without
let's just say it's in
well, yeah. I mean, even before, right? I mean, you have your scientists who are brilliant, but they know what they know. That automatically instills bias. But you brought up a very good point which is this notion that AI is going to wash away jobs. That may be the case in other industries, but absolutely not in our space. The need for, first of all, these are scientists. They don't trust. Anything they've gotta verify, they've gotta double verify, they gotta triple verify.
So whatever we do in silico on the computer is going to be wet verified. It's gotta be verified in a lab that, yes, this makes sense, I'm replicating this and therefore this gets me to my ultimate goal faster. So we are not seeing that at all. What we see is the advent of doing. A lot more, a lot faster. They have moonshot projects that have been on our whiteboard for two years and haven't moved. And you know, there's a sign that says Do not erase.
And you know, people have left that, but it's always stayed in the upper left hand corner. Now these things are starting to get pulled into view. These same folks, they're not losing their jobs by any stretch. They now get to work on. Four times the amount of projects than they ever had, so that's been very exciting.
That's cool. That's very cool. So, so this is maybe a good segue to our next topic, which is really around customers and customer acceptance, maybe even the customer experience. And I know that Ned, you're an expert in sales and business development and that's the role that you've played, um, with a number of companies really helping drive that customer and that value. Um, and I know that you're out talking to chemical companies and people across the value chain every day.
What are your customers excited and and or concerned about when they think about bringing in an AI based solution to their company?
I would say the, the folks who are looking inside the operational folks are worried about disruption, right? Transformation is scary. And so that, that's number one. Number two is do we really have the people in house that can leverage it? It's not just enough for a partner to hand this to us and run with it. We have to have the people that can run with it. And very often there is some change over there.
But I would say for the most part ultimately because it is new the finance folks can't really qualify it right? Or quantify it actually. Therefore, it becomes this, are we risk averse are we really ready for this? So my job as a business development person and my team is really there to help them understand what the business value is, right? The scientific value, I think is pretty demonstrable.
The economic value is really where the senior executives want to be able to sign off on it, but they're not necessarily willing to jump into the deep end of the pool without some, Either a reference or, you know, some business case built. So we spent a lot of time, you know, what, what would an acceleration of this project or. Financially for you what are the risks that you're seeing now from a supply chain?
We have one customer who had to take a product off the market for eight weeks because one of their small little chemicals. Yeah. Eight weeks is a major
That's a lot. Yeah.
a lot. I mean, so it's millions of dollars. And so when you have that and it's visceral like that is. You, you get to figure that out pretty quickly. But there are others that are just trying to figure this stuff out. business wise, it has to move a needle, right? We always talk about it's gotta save money, it's gotta make money, it's gotta de-risk, or it's gotta transform. If you can't do two of the four, then, you know, probably shouldn't do it.
Yeah, action then.
that exactly right.
Who usually brings you in? Where does that happen? Does that happen at. The, you know, at the the lab level, let's just say, or the product development guys, is it, is it the executive team that says, oh yeah, we know we need to do something different. Where do you see, how do you guys normally enter an organization? And then, we kind of touched on this, there's obviously different organizational priorities depending on where you sit and what you're looking at. How do you bridge those gaps?
Yeah, we just met with the CEO of a very, very large Fortune 1000, maybe even 500 CEO and his entire executive team, and they broke it out into four stages and research and development for a chemical company has a very tall pole in that tent, so is manufacturing and engineering. So they bucket their priorities based on. Revenue. Right? I mean, that's ultimately, especially if you're a publicly traded company, it's, it's, it's what moves the needle. So who brings us in?
To get back to your question, number one is very often it will be a product development. I. Manager, someone who is either behind the eight ball on their product development goals, right? Their product is delayed, it's over budget. Those are the folks who have the budget. But they're not the ones who can go and run these experiments. They're the ones who then have to bring us into the data science teams or the r and d teams very often. We need all three of those to get consensus.
It's a challenge in my world of, of selling and, and business development because you do cross all of these domains, right? If you're selling IT security, you get to sell to the security team and it's a pretty linear path for us. It cuts across all business lines. It cuts across manufacturing, it cuts across engineering, from how do we get from the bench to the market? either the r and d team brings us to product development. Product development brings us to the r and d teams, the two.
But nothing happens until we're speaking with senior executives because they're the folks who are looking outside the boat, as we say. And they're looking for icebergs. They're the ones who are saying, how can we get more gas into this engine? How can we do it without, without hiring a boatload of people without spending unbelievable amounts. Spending millions to make millions doesn't make sense. And so they're the folks who ultimately we need to get to, to really drive this.
Yeah, makes sense. And I mean, ultimately they hold the purse strings and make the big decisions. It also strikes me, Ned, that there is there's a real need for change management because this is a change in business processes, um, that have probably been in place, as you say, you know, maybe for a hundred years in some cases.
And it's a change that hits, interestingly, not just r and d and product development, but it's also a marketing effort and it's potentially a manufacturing and engineering effort. How do you see this playing out? The, the change management of introducing really a significant new tool and, and new approach to chemical innovation in a company.
Yeah, it's really interesting. Some of this is just absolutely institutional. So part of what we work through is are they culturally ready to really adopt a change. Now we do it. In a crawl, walk, run methodology. So we're not asking people to burn the boats and the bridges and adopt this new path. So we, we, we bring them along on this journey, but you're right, it's everything from administrative. How do we deal with people getting access to our systems? How do we give them access to our data?
Do we want to, who, how do we minimize that scope? All sorts of different processes. That's why this crawl, walk, run definitely works because they get the taste of it, they see the value of doing one or two projects, and then our customers almost across the board, add multiple use cases and models onto the platform. So you're a hundred percent right, especially in new technology. Those who are looking out of the boat tend to adopt earlier and make it happen, right?
Transformation is never, is never easy, and so it's hard. Exactly.
hard stuff, so that's great. What's, so, what's next for you? What should we be looking at for NobleAI in 24? What should we be looking at for AI and chemical innovation, uh, as we look ahead into 2024?
I think you're gonna see the rise of more and more partners adopting the specialized AI. For us it's just, going deeper and wider with our customers finding more opportunity to show
Yeah.
really, Victoria, there's no, there, there's nothing that we can't work on and move a needle on. If it has anything to do with science, we wanna, we want to at least take a shot at it. The second thing that we're really working on is expanding our presence through Microsoft. Microsoft is one of our lead investors. They say tremendous opportunity. Not only are they an investor, but they're also a customer.
They wanna bring this science-based AI or, you know, science AI for science to their customers. And so that's really been a, a starting to really take off. We're gonna be down at CERA week with them and very excited for that as well. So
We've come to the end of today's podcast. We hope you enjoyed your time with us and want to learn more. Simply visit TheChemicalShow. com for additional information and helpful resources. Join us again next time here on The Chemical Show with Victoria Meyer.
yeah, there's just a ton for us to do, but boy, there's, you know, right in our sweet spot. There's just a ton of customers to work with and ton of problems to solve.
Yeah. Cool. Awesome. Well, Ned, this has been great. Thank you for joining us today on The Chemical Show.
Absolutely. I really appreciate you inviting me, Victoria.
Yeah. I'm so glad to have you here and thank you everyone for listening. Keep listening, keep following, keep sharing, and we will talk again soon.