SPEAKER INTERVIEW
Sarah Eckersley, VP R&D, Industrial Intermediates & Infrastructure, Dow
Dr. Eckersley is the Vice President of R&D for Industrial Intermediates & Infrastructure (II&I) at Dow, which consists of three customer-oriented businesses of Polyurethanes & Construction Chemicals and Industrial Solutions. The II&I unit of Dow has over $16B in sales and provides specialty solutions in a wide range of markets such as consumer comfort, energy efficiency, transportation, and infrastructure. She will be speaking at Foam Expo North America and Adhesives & Bonding Expo on June 25 during the conference session “Harnessing AI and Automation for Material Discovery, Product Development Acceleration, and Performance Prediction”.
Why has the chemical industry increasingly adopted artificial intelligence and machine learning over the past few years?
The chemical industry has a long history of using artificial intelligence (AI) and machine learning (ML) in both operations and R&D. This has been largely focused on numerical-based data – for example, plant process data or research experimental data. That said, adoption has been accelerating because artificial intelligence and machine learning have become more accessible and affordable, thanks to the advances in computing power, data availability, and algorithm development.
Much of the recent enthusiasm is centered around the use of LLM (large language models). The exponential innovation in this space is driving the assessment and development of new use cases that have never been contemplated before.
Dow is concentrated on generating value from AI techniques for data-driven decision making.
As a customer-value-driven organization, our key focus internally and with our partnerships is to channel the possibilities on the tangible applications of AI that can transform our industry.
To date, Dow has implemented a wide array of AI capabilities:
• Greater visibility into the movement of shipments to improve transit time and reduce the overall delivery cycle, allowing Dow to proactively manage issues with inbound and outbound shipments that may affect customer order fulfillment commitments.
• We are well on our way to using Robotic Process Automation in supporting task automation, which will lead to automated workflows and leverage AI techniques for error/exception processing.
To keep our workforce prepared for digital and AI, we are partnering strategically with industry initiatives like DDMII (in manufacturing), University Partnership Initiatives (R&D and I/S), and key technology leaders like IBM and Microsoft to leverage technologies for next-generation capabilities that implement augmented-human decision making.
Do you think chemical producers in the US are enthusiastic about adopting these technologies, or do they exhibit reluctance?
There is likely a spectrum of answers to this question, which is correlated with the degree of readiness. Data is the fuel that propels AI / ML. Organizations that have a high degree of digital maturity (good data architecture and curation) are well prepared to adopt AI / ML and integrate with existing systems. Organizations that rely more on manual or isolated systems must first invest in the data infrastructure.
Data readiness also needs to be coupled with talent readiness, ranging from data engineers and scientists to visionary thinkers who can define new digital-enabled business models that take advantage of digital assets. The benefits and risks of AI/ML may not be clear or quantifiable, and chemical producers may face challenges in measuring and communicating the impact of AI/ML on their performance and profitability.
Lastly, the large language models have well publicized concerns related to ethical and social implications because of training-related bias and stereotypes, and this results in some caution.
• Dow’s strategy is to move from a chemical company that does digital to a digital company that does materials science.
• Digital is a company strategy, not just an IT strategy. It moves our ambition, and customer and employee experience forward to deliver growth for Dow and value for our stakeholders.
• Dow has invested more than $400 million in digitalization initiatives to accelerate the following strategic priorities:
• Materials science innovation and commercialization, including expediting product development using high throughput research, machine learning, and advanced modeling techniques, as well as quickening our speed-to-market through increased use of digital collaboration platforms and market listening.
• Providing a frictionless digital buying experience by infusing smart and responsive technologies, as well as extending customer reach with virtual experience platforms.
• Enhancing real-time manufacturing insights and operational data intelligence integrating multi-source data for intuitive decision making, increased safety, operational efficiency, and improving sustainability.
• We are tracking to deliver more than $300 million in incremental annual run rate EBITDA generation by year-end 2025, with an additional one-time $100 million improvement in structuring working capital efficiencies from these three focus areas alone.
• Dow’s digital strategy revolves around customers, employees, and the speed at which we work.
How does the predictive model capability serve as a tool for research and development (R&D)?
The process of product design and development – for the seemingly simple products we use every day – is an incredibly data and time-intensive undertaking.
For context:
• Time-to-market for a new-to-the-world chemistry could be 10 years.
• Time to develop a new product takes a little over 2 years.
• When we are optimizing a new commercial polyurethane insulation system, 36K data points may be collected when we are doing machine trials.
All that data is incredibly valuable – it enables us to make the right decisions. The inspiring opportunity is to harness all that data richness and transform it into knowledge wealth. As an organization, we have decades of experience formulating custom solutions which are used by our customers to develop their products. Our flagship digital capability, Predictive Intelligence enables us to focus all that historical knowledge on a single customer opportunity. All our processing models, lab data, heuristics, and knowledge can be used to predict a product for a customer, tailored to their specific needs, and get it to them faster. This condenses months of resource-heavy lab trials into seconds of digital data crunching.
But, it is a challenging and intensive journey for all the reasons outlined earlier:
• Acquisition of data that needs to be harmonized between teams around the world.
• Challenge of structuring data from disparate sources.
• Implementing new structures and changing the way people work.
What aspect of predictive modeling excites you the most?
Major investment is involved in developing a comprehensive digital tool like Predictive Intelligence. But the outcome is something that can transform the way the chemical industry works and collapse the standard timeline for realizing new product solutions for customers. Dow’s Predictive Intelligence AI models allow real-time prediction of formulation properties and can recommend formulations based on desired properties directly inputted by customers. For example, we applied predictive analytics to accelerate the R&D process for Polyurethane formulations by 200,000x — from 2-3 months to just 30 seconds. That excites me!
And finally, how does the utilization of technology help in aligning with customers' requirements?
Focusing on faster materials science innovation & commercialization: Dow partners with customers to innovate and bring the next generation of products to suit their increasingly complex needs for an increasingly complex world. In that innovation process, we infuse technologies like high throughput research, innovation centers, analytical sciences, and modeling which reduce commercialization times up to 2-3X.
Establishing end-to-end digitalization for supply chain resiliency, manufacturing insights, and operational intelligence: Our customers depend on our products, and they need to know exactly when a shipment will arrive, as it’s critical to their business. Knowing we are a reliable supplier builds trust and makes us a strategic partner. Dow formally established a Supply Chain Innovation team and launched the state-of-the-art Digital Fulfillment Center to accelerate the process from idea to implementation. Additionally, Dow’s Manufacturing 4.0 effort is focused on implementing digitalization to improve productivity and competitiveness, increase quality and reliability, and doing so even more safely.
Creating a frictionless e-commerce buying experience: Digital commerce meets at the intersection of the needs of our customers, employees, and the way we work to deliver value. Our evolution of Dow.com isn’t focused on building an experience that’s on par with our industry competition. We intend to be comparable with all industries when it comes to functionality and user experience, which means we’re continuing our path to make Dow.com look and function better than any company in any industry.
