Agentic AI As New Catalyst For The Chemical And Pharma Industries

Agent-Driven Procurement 

Strategic sourcing is a key lever for cost optimization in the chemical and pharmaceutical industries. Even the smallest deviations in forecasts or contract terms — sometimes down to decimal points — can have significant financial consequences.

An innovative pharmaceutical company with more than 500 million US dollars in annual purchasing volume and over 200 suppliers is currently exploring the use of AI Agents to build the procurement function of the future. An AI Agent is an autonomous system with a particular role that operates within a defined environment to achieve specific goals. Multiple Agents can work together on a task or pursue a shared objective. In other words, Agentic AI foreshadows a new era of virtual teams consisting of intelligent assistants for procurement.

Specialized AI Agents could soon generate demand forecasts, legal assessments, and bid analyses for buyers. Using natural language, these tasks can be delegated to the Agents, which then recommend concrete, actionable steps: “Increase glycerol inventory by 15% in Q1,” or “Activate clause 5.2 of framework agreement RV-1596 to call off additional volumes.”

Such intelligent Agentic teams could also transform negotiations, for example with suppliers. Negotiators could have trained negotiation Agents generate scripts, including a ready-to-read opening statement and arguments designed to secure concessions. A growing number of companies are already developing prototypes of such AI assistants — including in the automotive and energy sectors, driven by Agents’ transformative potential. 

Algorithmic Liquidity Planning

Liquidity planning in the chemical and pharmaceutical industries is a delicate balancing act with enormous financial implications. Hundreds of products, stakeholders, and locations must be aligned with a high degree of forecasting accuracy. At this level of complexity, manual Excel-based solutions quickly reach their limits. Yet reliable forward visibility is critical for investment decisions, solvency, and financing costs.

BASF addresses this challenge with its Algorithm-Based Cash Forecast (ABC Forecast), an AI-powered solution that generates a rolling six-month liquidity forecast every month. The project began in 2017 with around eighteen months of intensive groundwork: data preparation. Historical SAP data and external economic data were collected, analyzed, and adjusted for special effects.

Initially, the ABC Forecast learned alongside human planners and has since become an integral part of BASF’s financial planning processes. Depending on the legal entity, different mathematical and statistical methods are applied. Today, the ABC Forecast covers around 60 percent of the BASF Group’s cash flows. As a result, planning at the chemical giant has become not only more accurate, but also significantly faster. BASF’s ABC Forecast vividly demonstrates the potential of AI for financial planning in the industry.

Chatting With The Lab

In specialty chemicals, finding the right solution for customers and suppliers can take weeks or months, even when suitable products may already exist. Especially in complex applications such as the coatings industry, every detour in the search process costs time and, ultimately, money.

Evonik addresses this challenge with Coatino and AI-powered product recommendations. Coatino is a publicly accessible digital lab assistant for the coatings industry. It combines natural language processing (NLP) and machine learning in a user-friendly interface to “understand” additives and deliver precise product recommendations.

Via a web interface, customers can submit queries in natural language, for example about additives for defoaming or surface hardening. Essentially, they are “chatting” with Evonik’s lab. Around 1,900 reference formulations and more than 200 product properties are continuously enriched by up to 120 new test results per day from a lab in Essen, which Coatino uses to learn. The result: shorter manual search times, higher customer satisfaction, and more targeted sales steering through better-matched product recommendations.

Similar chat-based applications that make it easy to access corporate knowledge have already proven their value in other areas. How they work and what makes them special is explained in our blog post about “chatting with tables”.

If you want to learn more about Agentic AI and its possible applications, we also recommend our whitepaper “The Agentic Shift.” It examines the promises, limitations, technical architectures, and real-world use cases of Agentic AI, and offers a glimpse into the future by asking where this “Agentic journey” might take us.

Conclusion: Rethinking Chemicals and Pharmaceuticals with AI 

At the beginning of the 20th century, Fritz Haber and Carl Bosch worked on the idea of synthesizing ammonia from nitrogen and hydrogen under extreme pressure and heat. Many dismissed it as “too risky, too dangerous, too expensive.” But Haber and Bosch persevered, built prototypes, won over the public, and ultimately found the catalyst that made the reaction possible. Today, the Haber–Bosch process underpins food production for large parts of the world’s population. Their idea was “transformative,” in today’s terms.

Viewed this way, Artificial Intelligence is nothing new to the chemical industry. It is another technology that the sector can successfully integrate over time. Each of the cases described uses AI to solve problems that occur hundreds of times in Germany alone. All of them started with vague ideas and many open questions. Today they are shaping the future of chemistry. Those who invest wisely and build experience now will have a clear advantage tomorrow. One thing is certain: the potential of AI is far from exhausted.

Are you looking to use AI in your organization to create tangible value, or are you already running initial applications and ready to take the next step? Feel free to get in touch for an informal initial conversation.