What AI can do for supply chain management in speciality chemicals
Combinations of machine learning, generative AI and agentic AI will provide supply chain managers the tools to manage today’ s complex supply chains. But first, you need to get your data harmonised, says Dr Marko Lange, SAP’ s global head of chemicals
Speciality chemicals supply chain experts don’ t have it easy. Industry overcapacity, anaemic demand, high energy and feedstock costs, and – foremost on many minds – the impacts of tariffs count among today’ s challenges. It would be naive to assume that tomorrow’ s hurdles, however different they may be, will be fewer or lower.
The operating environment in the speciality chemicals industry has always been complex, but there was an underlying consistency to it. Now we seem to find predictability only through the regularity of surprises. The emergency of generative AI( genAI) and, fast on its heels, autonomous generative AI agents( agentic AI), represents two such surprises for the speciality chemicals supply chain. They should be welcome ones.
A synergistic future
AI is increasingly used in supply chain optimisation by many companies, and adoption is in full swing. Industrystandard systems take into account supply chain-specific variables, such as predicted demand, stock levels of raw, intermediate and finished materials, and transportation costs. However, that is not enough anymore.
Commercially relevant variables, such as seesawing raw-material prices, country- and region-specific risk profiles based on geopolitical uncertainties, environmental regulations and tariffs, are not new( although the current tariff- related uncertainty probably does qualify). What really is new is that, with a combination of generative AI and machine learning, you can incorporate such variables and many others to make better decisions, faster.
The AI tools are evolving quickly, but there has been solid progress in a few areas already. GenAI lets users run scenarios and query using natural language, with answers returning in natural-language form where appropriate. This improves the usability of these systems for experienced hands and speeds the learning curve for new users, while obviating the need to involve technical staff for custom queries
Master data integration and checking may not seem worldchanging but it is a great example of AI addressing what can be a tedious issue for human staff. Supply chain planning systems depend heavily on master data and better data means better planning outcomes. AI quickly roots out data problems and suggests fixes.
AI is enabling the incorporation of external signals such as the commercially relevant variables noted above to sharpen both shortand long-term demand forecasts, thus aiding demand planning. It also improves inventory and supply planning by combining factors like target inventory levels, supplier reliability, demand volatility, lead-time variability, constraints in production capacity and transpor, t and an ability to simulate demand surges, supplier shutdowns, tariff changes, port congestion or other issues.
Finally, AI can also help sustain and grow sales. The higher-precision
58 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981