Alan Robinson, commercial vice president for North America at Dow Polyurethanes, details his company's new journey into the world of predictive intelligence and shares insight on what it could mean for end users within protective coatings and corrosion control. Other topics discussed on the podcast include industry trends; potential implications for customers; and lessons learned from Dow's growing emphasis on digitalization.
See below for a complete transcript of the recent episode. For more information, visit www.dow.com/polyurethanes.
Ben DuBose: Alan, good morning. How are you?
Alan Robinson: Good, and good morning to you as well, Ben. Thanks for having me on.
BD: No problem. Glad to have you. Before we get too deep into the podcast, I’ll start by giving you a chance to briefly give a biography of both yourself and the company. For any of our followers that might not be fully familiar, just fill us in about what the expertise is that you have, or your company’s expertise.
AR: Absolutely, Ben. Why don’t I start first with just the company, Dow Inc., a former Dow Chemical company. We specialize in making multiple different materials, from plastics to wet chemicals, industrial-based chemicals, and today what we’ll really talk about more is in the polyurethane space of chemicals that we offer to various industries and segments.
If I spend a little bit of time just on me and what I do, I’m the North American commercial vice president for polyurethanes portfolio. What that means, think of it as the general manager running the North American business. But the capabilities and the things we talk about today are truly global. Within Dow, I wear multiple hats. That’s one of my jobs. The other job is really leading, globally, the strategy around some of our high-value polyurethane systems business, as well as being one of the key drivers in our digital transformation within the business and within Dow itself.
What we talk about today, as you’ve been seeing in the media and different resources, is we’ve launched a global capability that we’re calling predictive intelligence, which really gets at formulating expertise and making it easier for our customers to do business with us and making it faster, more efficient for them to do business with us.
BD: As many of you have probably surmised by now, Dow Polyurethanes, as Alan was explaining, represents the polyurethanes business within Dow. As far as polyurethanes go, that’s a big part of the equation for our audiences at NACE as well. Obviously, they have technical properties that can be used for corrosion resistance. For example, we often write about contractors that will apply polyurethane-based coating systems.
With that as the backdrop, I’m going to read the first paragraph of a press release that Dow Polyurethanes recently put out. Alan was just referencing this. You can find it at www.coatingspromag.com or www.materialsperformance.com. Then I’m going to get Alan to talk a little bit more about what it means in greater detail, as far as the implications for the industry. Here’s the lead:
“Dow Polyurethanes, a business division of Dow, is accelerating its digital transformation through increased integration into artificial intelligence (AI) and predictive capabilities geared to increasing its digital IQ across all business operations. Dow is collaborating with Microsoft to accelerate the integration of machine learning and AI solutions into a novel predictive intelligence capability that will transform product development process and enhance customer collaboration.”
In short, Alan, what does predictive intelligence mean as far as the industry is concerned? What are the tools that this can give your business or any business potentially?
AR: Think of it as a couple different stages, Ben. When I get back to “this is all about speeding up and making it easier to do business with our customers,” think of it as a few stages. So if we simplify this. Stage 1, we do business all around the world. Let’s use your audience as a coatings, for example. We could pick a favorite coating. Maybe it’s the garage or group of coating application. All around the world, we do different formulations, different testing, different data that we might have. To go into one of those formulations, it could be 15 different variables. Within each of those variables could be thousands of options. Then just the percentage of that formulation can change between those 15 different variables.
When you simplify this, think of it as a design of experiment that is millions, and it just becomes too much for the human brain to do. We do this today in the polyurethane industry. The discovery phase, the selling cycle, it could be anywhere of 18 months from customer raises a problem to you go into the bench in your lab and you start to formulate all these things, and then you try to trial, maybe you fail, you go back, you tweak it again. So this could be an 18-month cycle. Some are longer, some are shorter. Now think about Stage 1 as just tapping into all that knowledge that we already have. What would normally take us in discovery phase of what could be two, three months now happens in a matter of 30 seconds. We would type in key performance indicators that the customer is looking for, and we’re able to harvest all of our big data that we’ve been working on and building up, and come up with something that we’ve done today or have done in the past that would be a good starting point formulation to work in that.
But then you've got to go over into Stage 2. Stage 2 is oftentimes we don’t have solutions or we may not have done this before. Trends are evolving. There’s a new chemical that’s being tested for waterproofing or for chemical resistance. Oftentimes you may not have a formulation sitting on the shelf. Well, now introduce the Microsoft, the machine learning, artificial intelligence capabilities that can help predict formulations that maybe we haven’t put together or not currently in our “catalog.” Now it will give us, based on all the testing that we’ve done, all the data that will upload into this capability, it will give us a projection of what the formulation should look like, given that we’ve never done it. This is a great opportunity to really accelerate our innovation pipeline and also accelerate doing business with customers and solving their problems.
Stage 3, although it may not be applicable in some of the areas that you do business in, as I think of formulators and applicators applying it down, but in many cases in polyurethane you have to marry what chemistry with equipment. There’s just as many variables in the equipment. Being able to build machine learning on equipment recommendations, equipment settings, how it marries with the chemistry is going to be where we get to in the future as well. When you think about this, the attempt is to take a sales cycle that is exceptionally long today and be able to simulation everything to shrink that sales cycle down, cut it in half, more than 50%, making it easier, more effective, and more efficient to do business with Dow and to meet our customer needs faster.
BD: What type of trends have you seen in the industry as far as the increased digitalization? Just on a very broad level. Why is this the right way to go? What type of feedback have you gotten from the customers that you deal with that lends you to think that the industry is ready for this and that this is the right direction to go in?
AR: Let’s first start with the trends. On the trends side, I think there’s been a lot of activity in digitalizing supply chains. You often like to use other industries. Main players like a Home Depot or a Wal-Mart that have really taken digital supply chain to the next level. Or even Amazon, where connecting supply chains and ERP systems between different companies to really make it digital, make it more effective.
There’s many of those trends here in this space as well, where you're looking at digitizing your entire “end-to-end.” I produce and I can ship to a customer, and making that digital. Dow’s also doing that as well, ensuring that we can digitize the end-to-end process. We see digital starting to take place in troubleshooting, whether that be sensors on your customer’s equipment and doing more digital troubleshooting from afar. That’s an exceptional way to also help address and do troubleshooting for customers. Digitizing your R&D process.
When you think of a polyurethane system, for instance, you're customizing each formulation for every customer. No one product is sold to two customers. It’s a very highly customized solution for all of our customers that are our there. Digitizing the whole formulation and customization process will help speed this up. That’s why we looked at this going, “We can all do the digitalization of end-to-end, and we can do the troubleshooting, but we believe there’s a tremendous amount of value in shortening the sales cycle to meet customer solution.” We think that’s a winning proposition.
BD: Have you guys had any feedback, or maybe lessons learned I think is what I’m going for, for other people or companies in the industry that might be looking at this and considering perhaps a similar transformation? What type of experiences have you had in the early going that might be useful for other people in the industry to learn from?
AR: It’s a great question, Ben. What I’ll tell folks is, one, be patient and know it’s a journey. You don’t get there overnight. Some of the key things we’ve learned is — take our space in polyurethanes — you really have to break it down into simple, basic steps and focus your area on certain segments. You can’t take the shotgun approach and do it everywhere. We’ve really isolated it down to say, “Here’s certain areas that we know we’ll be more successful in.” Getting it standing up first and then work our way through the various different applications that we can build the model, the capabilities to get there.
It is a journey. This is not something you raise your hand and say within a couple months you're going to be there. This is a multi-year journey that you’ll take to really digitize the whole process to go after it. You have to stay focused. You have to be willing to prioritize and stay away from some other initiatives that might veer you off of this attention, and you have to realize that you’re probably going to need some help. You won’t have all the capabilities internally.
BD: Here at NACE, and I mentioned this earlier, our audiences are largely interested in polyurethanes as they pertain to things like coatings, linings, insulation, and all these other technologies that can be used for corrosion control. Are there any examples that you can give as far as the types of trends or information that can be predicted within material science? Just talk, if you could, about some of the properties within that field that you might be able to predict with this type of machine learning, predictive intelligence.
AR: Certainly. We won’t be able to go into a tremendous amount of detail, but I can talk at a high level as you think around some of the different areas. In the CASE space, is really what you're referring to, we’ll call industrial or CASE — the coating, adhesive, sealant, elastomer side of the business in polyurethanes.
You can look at — we often talk about KPIs, key performance indicators that folks would be looking for. Whether that’s bonding material or a substrate that you're going to be bonding to, or what kind of chemical resistance, what kind of waterproofing resistance that you might be looking for. The problem with trying to answer in the CASE space is so highly fragmented. There are so many different applications. There’s so many different KPIs. I’m trying to give you a high level of what you would focus in on. If you start to — you have to figure out what the key initiatives are in each of those various applications and segments. I’ll give you an example. I know you mentioned insulation in one space in your writeup. Folks might be looking for insulation value, they might be looking for how to fill a cavity, they might be looking for density, they might be looking for sustainability, features.
All of those KPIs can be plugged in to figure out what formulation best meets the needs for a customer. Similar thing as we build out our CASE initiatives in those spaces is we’ll have the same thing, whether it’s substrate bonding, whether it’s waterproofing, chemical resistance, or you get all the way down in some of the elastomer areas of hardness and types of materials that we’re dealing with. So all the key performance indicators. The intent is if you simply this, if a customer has a problem, they identify what the problem is and we’re able to find a solution that matches up with that. And if we haven’t done it before, the Microsoft capability will help us find the formulation that could do it.
BD: As we wrap up the podcast, what’s the timetable moving forward with all of this? What are the next steps on the horizon as Dow Polyurethanes continues to roll all of this out?
AR: If we look at end of the year, we’ll have capabilities in a particular segment. We’re focusing in on — the first segment to launch a truly predictive capability is going to be in what we call a consumer science space. That capability — we’ll have internal capabilities by the end of the year. We’ll begin customer trial, customer invites in the first part of next year. By the end of 2021, what we’ll be launching also is a website available to start utilizing these capabilities from an external standpoint. So a customer portal to go in and start identifying what solutions could match needs there. That’s our intent.
Phase 1, we’re already there. I mentioned the three different phrases. Phase 2, we’ll be there in a particular segment by end of the year, with the intent to extrapolate that out over multiple segments, with the target to be there across all segments within a two-year timeframe. Like I said, it’s a journey, Ben. It doesn’t happen in just a couple months. Then like I said, we’ll open up customer portals in 2021 to begin and start using and utilizing these tools. And our internal team will start being able to utilize the tools even deeper. So Stage 2, 2021.
BD: Alan, before you go, for any of our listeners that want to keep track of your journey at Dow or get more information, what’s the best way that they can do that? I’m guessing it’s the website, but basically plug anything that you want to plug that might be of value to our listeners.
AR: Absolutely. First and foremost, www.dow.com. Our website’s been revamped. That’s where our predictive intelligence will also be sitting, is on our dow.com website. I would say that’s the best place to go to to really start tracking progress, raise your hand for questions, start checking in on what’s happening predictive intelligence-wise. The other thing is always keep an eye on LinkedIn. We’ll be sending out updates. We’ll be using that forum to provide any information, updates, and access to folks that may have questions. You can always reach out in that space as well, on LinkedIn.
BD: Sounds great.