Dow Inc. uses AI to create new products

From softer mattresses and more supportive sneaker soles to quieter cars and more efficient refrigerators, countless objects that make up daily life are enhanced by polyurethane chemistry. Dow Inc., a 123-year-old materials science company based in Midland, Michigan, creates the materials that make up these products through its Dow Polyurethanes division. The materials are designed to order for customers at the company’s 106 manufacturing sites across 31 countries.

“No two customers buy the same product,” says David Parrillo, Vice President of Research and Development at Dow. “That means we have to formulate and customize exactly for what each customer needs.” In step with the company as a whole, this formulation process is undergoing a digital transformation.

“Our overall goal is to provide a seamless and enhanced digital experience for the customer along with our products and services,” Parrillo says. Dow is one of the three biggest materials science companies in the world, and its products, which include plastics, industrial chemicals, silicones, and polyurethanes, touch nearly every part of every life on the planet.

With a goal of becoming the most innovative, customer-centric, inclusive materials science company in the world, Dow adopted all three Microsoft clouds: Azure, Office 365, and Dynamics 365. “From research and development to supply chain to order management to manufacturing, many steps and processes across Dow are being digitalized today to enhance our digital IQ across all of the company,” Parrillo says.

The improvements brought by this sweeping digital transformation include greater supply chain visibility and resilience, smoother communication and collaboration among the company’s 35,700 employees worldwide, and faster, more actionable insights on the data generated across all its business processes. Among the advances made possible by digital transformation, the polyurethane formulation process is a standout example of how Azure AI and machine learning can fundamentally transform a legacy process to bring greater value to customers.

Reducing records analysis from months to seconds

When a Dow customer needs a new polyurethane formulation to build or enhance one of its products, the development from start to finish has traditionally taken about 18 months. For example, a customer might approach Dow looking for a polyurethane formulation to create a particularly soft mattress that maintains its support. Before machine learning, this would be the beginning of a long, manual process. Scientists would sit down at the laboratory bench and dig through textbooks and archived records of past work to guide their experimentation to find a new formulation.

With Azure Machine Learning, all the team’s prior knowledge, expertise, and records about previous flexible foam formulations are fed into machine learning AI models. The models are then able to take that knowledge and predict where there are gaps that could suggest possible novel flexible foams. Rather than taking four to six months of work by multiple experts, the algorithms can sort through millions of possible combinations and suggest promising areas for experimentation—in seconds.

“We have 123 years of history of going into the lab and breaking out the beakers to solve an issue,” explains Alan Robinson, Commercial Vice President at Dow. “Whereas now, instead of the beakers, we're going to the laptops and the keyboard and the mouse and, with a couple of clicks, coming up with a solution almost instantaneously to solve our customer needs. It’s a mindset shift.” That shift is being accomplished through an integrated partnership with Microsoft.

Combining chemistry and technology

Dow is participating in a multi-year strategic AI and sustainability partnership with Microsoft. This partnership will deliver breakthrough AI innovation through R&D-to-R&D collaboration and help Dow meet its Sustainability 2025 Goals. To get the most out of the technology they use, Dow empowers its people to work directly with Microsoft experts.

Gareth Bland is a Senior Data Delivery Scientist at Microsoft who has collaborated closely with Dow chemists and data scientists. “The power of the cooperation between Microsoft and Dow Polyurethanes comes from the ability to combine our knowledge from the Microsoft Azure AI and machine learning side with Dow’s scientific knowledge,” Bland says.

“Among all these scientists, you have hundreds of years of [cumulative] expertise in chemistry, and now you're bringing in all these digital tools,” Bland says. “So the role of the data scientists becomes a kind of translator between, ‘This is how we think about chemistry at Dow, and this is how we can leverage that through machine learning.’”

Parrillo concurs. “The key ingredient to this partnership between Dow and Microsoft has been collaboration,” he says. “At Dow, we know materials science, and Microsoft has enhanced capabilities for machine learning, AI, and the platforms that they reside upon. Bringing all of this together and allowing the diverse perspectives from both of our companies to create a product like predictive intelligence has enhanced our capabilities at Dow.”

Anticipating customers’ needs

In the future, Azure AI and machine learning will help Dow Polyurethanes build on this accelerated formulation process to develop new products before corporate customers even know they need them.

“One of the interesting ideas from the formulation process with Dow Polyurethanes is you can start to predict new products that customers are likely to come to Dow for,” explains Bland. “AI and machine learning can identify a product that has yet not been made but is likely to be needed in the future.”

This process has been rolled out in Dow’s ComfortScience™ division, which includes materials used in products including mattresses, yoga mats, bicycle helmets, and other consumer goods. “Our plan is to roll this out across other segments, completing it by 2022, where we will have predictive intelligence capabilities based on Azure AI and machine learning across our entire portfolio,” says Robinson.

Adoption of Azure AI and machine learning is one key part of Dow’s ongoing initiative to lead the industry in providing high-performance products that meet critical market needs and contribute to a more comfortable, healthier, and more resilient society.