Predicting the future with digitalization

What if we could predict our customers’ needs and create solutions for them?

Diving further into our mission to solve our customers’ pain points before they even arise, we created the Predictive Intelligence (PI) capability to do just that. The PI capability radically improves the product formulation process by predicting formulation properties and simulating customer processes. 

Using the power of machine learning and Artificial Intelligence (AI), in combination with Dow’s material sciences expertise, the PI capability accelerates innovation, and will ultimately improve the customer experience by predicting trends and creating new products at a faster pace than ever before.

Predicting success

At Dow Polyurethanes, we are constantly designing different formulations for our customers all around the world. If we look at a single formulation, there may be over 15 variables that need to be considered, and for each of these variables, we might be dealing with more than a thousand different options.

This means that in the design and experimentation phase, we could be processing over a million different data points. This is a huge amount of data to work through, so normally the discovery phase of product development alone takes 2-3 months. With PI, we can take those 2-3 months down to approximately 30 seconds, speeding the process up by around 200,000x.

Making the process faster, easier and more efficient for customers by using the knowledge we already have and feeding it into our new capability, is just the first phase of what we plan to achieve with PI. In the second phase, we plan to use PI to predict formulations that do not currently exist, using the power of AI and big data to predict trends and project new formulations.

Our eventual goal will be delivered in phase 3, when we make the leap from pure chemistry to how it functions with equipment. Just as with formulations, there are many variables and data points to consider in equipment and we hope to be able to use machine learning to generate equipment recommendations and settings.

Collaborating for impact

In the all-encompassing digitalization journey we’ve undertaken at Dow Polyurethanes, we quickly realized the value of collaboration – both internal and external. Internally, this meant collaborating across functions to increase our digital IQ across all business operations in a cohesive way. We also realized that it is impossible for a company to have all capabilities present internally and that we owed it to our customers to find the best solution possible, wherever that was.

That’s what led to our collaboration with Microsoft. We managed to combine the best of both worlds by bringing together Dow’s expertise in materials science, reaction kinetics and rheology with Microsoft’s expertise in AI machine learning leading to our flagship PI capability. By looking for the best partner in this project, we were able to make use of Microsoft’s prized Azure Machine Learning and eventually create a better product for our customer than we would have been able to on our own.

With customer satisfaction being at the heart of why we developed PI to begin with we also wanted to make sure that their input is captured in how we bring the tool to them. This is why we are collaborating with select customers and soliciting their feedback to develop a web portal where they will be able to navigate the entire process from product search to final delivery in a single interface. Customer trials will begin in early 2021, with the web portal becoming available externally, mid-2021.

Our digitalization journey has only just begun, with much more planned not only for PI but with other transformative initiatives as well. Learn more about the Dow Polyurethanes’ digitalization journey as we attempt to transform the industry and how we work together.