At SWSCC Suppliers’ Day, Bill Birt, a principal engineer in R&D focused on cosmetic process development, will present “Smarter Decisions Through Data and AI in Cosmetics.”

His presentation will examine practical applications of data and AI in cosmetics across R&D, quality, operations and supplier contexts. In it, he argues that AI amplifies—rather than replaces—technical judgment when historical lab and process data are treated as strategic assets.

Using the historical analogy of the power loom, Birt’s talk frames technology as a force that reshapes work while expanding the value of skilled expertise. A central case study describes predictive stability modeling built on approximately 20 years of freeze–thaw data. After extensive cleansing to address inconsistent terminology, missing values, and evolving methods, binary pass/fail models (developed in RapidMiner) provided early risk signals that helped prioritize physical testing for new formulas. The work underscores that data standardization, not algorithm selection, was the primary bottleneck, and that purely historical outcome labels have inherent limits for truly novel chemistries without molecular-level features.

Beyond the laboratory, the presentation highlights process dashboards that convert manufacturing signals into earlier visibility of drift, enabling proactive intervention while remaining complementary to existing controls. Generative AI is positioned as an investigation partner that organizes prior knowledge, surfaces root-cause hypotheses, proposes alternatives, and structures follow-up experiments—provided prompts are carefully engineered with role, context, evidence, desired output format and constraints, according to SWSCC.