Hannah Melia of Citrine shows how AI can address some of the commonest supply chain issues in speciality chemicals
Overcoming supply chain disruptions with AI-driven reformulation
Hannah Melia of Citrine shows how AI can address some of the commonest supply chain issues in speciality chemicals
Citrine is an enterprise software company providing artificial intelligence( AI) and machine learning( ML) tools to the chemicals and materials industries, particularly for product R & D. Unlike most others in the field, it is built around‘ small data’, because it understands clients are not in a position to provide data on millions of past experiments.
Data is often confidential and, in any case, collecting it is expensive and time-consuming. Most Citrine clients are working with only a few dozens of data points to train their AI and ML assets in the platform. This enables them to get going quickly and innovate to get new products to market faster or to build more resilient supply chains.
In chemical development, there are challenges, starting with he constant tension between cost and performance. How do we reduce costs without sacrificing some element of the performance customers expect? At the same time, how can we be faster when the customer wants our product right now? On a wider note, how can we bring in sustainable ingredients?
There is also the regulatory landscape. Sometimes ingredients that have been used for decades are suddenly not allowed in certain regions. Moreover, the chemicals industry is ageing. People are retiring and often taking their knowledge with them when they go. On top of all that comes supply chain disruption, which can be highly unpredictable and can make or break companies.
Supply chain disruptions
Supply chain disruptions— whether caused by tariffs, natural disasters
ORIGINAL INGREDIENTS CATALOGUE
ALTERNATIVE INGREDIENTS CATALOGUE
•
• •
FILTER BY COST
POSSIBLE INGREDIENTS
EXPLORE USING Al or market volatility— demand swift and strategic responses from speciality chemical companies. AI platforms empower organisations to adapt to such disruptions with agility. Companies creating chemical products must formulate or synthesise products that meet a variety of objectives. Raw materials are meticulously chosen to optimise technical performance, sustainability and business requirements.
However, in an increasingly complex world, it is becoming more likely
200-
CXl '-
�
I... a,
0 I...
150
0
· 0-C: 0
.c:
0 100- a,
50-
I 50
SET OF Al GENERATED RECIPES READY FOR TESTING
INGREDIENTS RESIN COLORANT
ADDITIVE
Figure 1 – Understanding how different chemical approaches can hit the same performance targets.
Figure 2 – Generating a bank of recipes using AI
r
•• Possible formulation types
••
Performance target
Several formulation types meet target
•
that an ingredient may become unavailable. The worst-case scenario is a halt in production and a potential loss of market share as disappointed customers seek alternative suppliers. Below are outlined three ways in which AI, or more specifically, ML for chemical formulation, can be used to mitigate the effects of supply disruption:
• Having a plan B( or multiple sourcing)
• Local sourcing
• Speed
'\
•-
•
•.• ". •
• ••
• •,
• •
•
•
Mechanical property A
• •• •• •
• •
� ®·..,..
Figure 3 – Evaluating performance targets vs. product types
64 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981