The strategic role and impact of AI
Across industries and applications, AI adoption is accelerating. In pharmaceuticals, for example, AI has reduced drug discovery timelines by up to 50%, improved clinical trial design and enabled dozens of new therapies.⁵ Pharmaceutical R&D is in many ways well-suited to AI tools. Many pharmaceutical products have few active ingredients (or only one) with specific biomolecular target(s). AI models have been successfully deployed (alongside high-throughput testing) to dramatically accelerate drug discovery and commercialization. The chemicals industry faces much greater complexity, as products often require balancing dozens of ingredients and performance variables with intricate regulatory and sustainability requirements.
Even so, AI is proving to be a competitive differentiator that transforms the economics of R&D. The traditional R&D cycle in chemicals is often slow and costly, hampered by manual experimentation and fragmented data, but AI is transforming this paradigm. Platforms such as Citrine Informatics enable researchers to screen hundreds of millions of candidate materials and formulations, narrowing down to those that meet both technical and business criteria and achieving up to 80% reductions in R&D time.⁶ The predictive capabilities in their AI models mean that fewer experiments are needed to successfully meet chemical and material property requirements and performance targets. Researchers can screen broadly but select fewer candidates for physical testing with a higher likelihood of success, reducing costly and time-consuming experimentation. In one notable case, a customer improved EBITDA by 50 basis points in less than two months by reformulating products to optimize cost and performance.⁶ New product development and qualification often require substantial testing in multiple settings, including both laboratories and pilot plants. In addition to accelerating experimental design and execution in the laboratory and pilot plant, AI is being used to facilitate supporting processes and systems. Chemical companies are combining data visualization, feasibility review, sample request and approval, and other tasks into unified applications and streamlined workflows, reducing error and saving significant time on non-value-added administrative tasks.
Beyond efficiency, AI is revolutionizing how organizations manage and share knowledge. AI-powered research assistants like CHEMY LANE enable users to ask questions in natural language, receive sourced and documented answers, and automatically extract actionable data from scientific literature. The Citrine Platform’s Catalyst digital assistant leverages large language models to streamline research processes, summarize disparate data and facilitate “scientist-to-scientist” interactions. AI can quickly consolidate and interpret data from multiple sources, transforming fragmented information into actionable insights. In large, siloed R&D organizations, this capability improves product development cycles, reduces time and cost and enhances the quality of offerings. AI-driven automation in regulatory submission and content writing has reduced report generation time from weeks to minutes, streamlining compliance and enabling faster product launches. Furthermore, AI-driven portfolio management tools provide real-time risk modeling and scenario analysis, enabling executives to make better investment decisions and improve return on research capital (RORC). These capabilities translate into outsized gross profit, improved return on invested capital (ROIC) and sustainable competitive advantage.
One noteworthy development is the emerging use of neurosymbolic AI (NSAI), a hybrid architecture that merges the statistical learning capabilities of large language models (LLMs) with the structured logic of symbolic reasoning. Critically, this addresses one of the shortcomings of LLMs, which is their lack of understanding of causality. The potential of NSAI models is rooted in their inclusion of clear causal relationships, making them more than just black-box text prediction algorithms.⁷ Because they incorporate logic and processes, NSAI models can be configured to reflect the actual decision frameworks of an enterprise, enabling insights that are not only grounded in causal intelligence but also explainable, repeatable and aligned to business strategy. EY Growth Platforms (EYGP) is helping companies deploy NSAI models to extend beyond investments in generative AI to unlock deeper insights. The NSAI platform resides within the client’s data environment and unifies enterprise data with over 110 million external data sources to build proprietary workflows. EYGP is working with companies in life sciences, consumer products and specialty chemicals, among other sectors, to tailor R&D priorities based on real-time consumer feedback and scientific literature, accelerating the alignment between market needs and scientific innovation.⁸
Finally, sustainability is also being redefined by AI. In a recent survey, 70% of chemical manufacturers cited AI as critical to their sustainability strategies.⁹ The value of sustainability-related company transactions in chemicals reached $14.7 billion in 2023. Companies such as PPG have set ambitious goals to reach 50% of sales from sustainably advantaged products by 2030, up from 39% currently.¹⁰
Case studies and implementation considerations
Case studies: AI in action
