FEATURE Smart factories & software
THE AI ADVANTAGE IN CHEMICALS
Octave examines how AI will have a positive impact on the role of workers in the chemicals and pharmaceuticals industries
AI will not replace the chemical sector’s employees. Retirement will. But AI will help eliminate the ‘knowledge traps’, and make information accessible to younger workers
task mix within jobs rather than removing the jobs themselves. In a report by the World Economic Forum
R
arely has the introduction of a new technology been accompanied with doom- mongering from the very
people who were selling it as was the case for generative AI: In May 2025 Anthropic’s chief executive Dario Amodei told Axios that AI could eliminate half of all entry-level white-collar jobs within five years and drive unemployment to between 10 and 20 percent. On the industry side, predictions have been less dramatic but still promising a major shift: ING believes that AI adoption could lift annual productivity growth by up to 3%, reversing two decades of near-flat performance.
That perspective is an attractive one for
Europe’s embattled chemical sector. It has struggled for years with higher labour costs than its competitors and labour productivity that has decreased by 25% in the past decade. That fall reflects external factors like reduced plant utilisation, lower investment and Europe’s energy and feedstock disadvantage, not workers suddenly becoming less efficient. The context has pushed some companies to jump the gun and tie today’s workforce reduction programs to the hopes of billion- dollar productivity gains from AI in the future. But what does a sensible hiring strategy look like in the AI age? First observation, the pace of AI deployment is not in doubt. A survey of German chemical and pharmaceutical companies found that 76% of respondents were using AI in 2025, compared with 34% in 2020. Surveys generally suggest that the top use cases,
20 October 2026 | Automation
process optimisation, quality inspection and scheduling, have been embraced by one in two companies.
These choices reflect an evolution: after two years of standalone pilots that often stalled, there is a clear trend towards embedding AI in the tools and workflows people already run to try to scale it faster across multiple plants. That includes process simulation software, Enterprise asset management and asset performance management systems, QMS and digital twins of the plant, to name a few. That gives an indication of the tasks that can indeed be rationalised, often to the effect of requiring fewer people: organising engineering documentation, compiling reporting and data analysis, spending time on manual and reactive procurement processes or scheduling by hand, among other administrative work. A survey carried out by Octave last year found that operations teams tended to lose an average 117 work days a year to manual reporting and data consolidation. That is more than half of one person’s annual capacity. It should be noted, however, this potential for optimisation predates generative AI, and many large players have explored the potential productivity gains brought by digital twins or EAM systems, for example. That notion of “redeployment” is an essential one because the idea that you might need fewer people to run a chemical plant might not materialise at all. Major workforce forecasts do not suggest AI will simply wipe out the core operational roles in chemical plants but offer a more nuanced picture, with technology shifting the
projecting job growth and displacement from AI by 2030 finds no chemical-plant role set to shrink because of AI. It even sees operator positions rising by 11%, with white-collar roles like chemical engineer holding broadly stable. What this shift in the human-machine frontier looks like is already visible in many plants, with productive and financially rewarding AI use cases falling mainly into three areas. First, absorbing and contextualising vast amounts of information, second, detecting patterns and performing analyses, and third, advancing and packaging the work itself, for example by drafting a work order or proposing an update to an asset record, while leaving consequential decisions and execution with a person. This is the familiar “human-in-the-loop” model. All these use cases have clear financial value but do not necessarily mean operating with fewer workers. They mean workers spend a greater share of their time making decisions and do so with more context. There is also a notion called the Jevons Paradox. When a technology like AI makes a task cheaper and faster, organisations often do more of that task, rather than needing fewer people.
This could transform an industry where
around 30% of employees are due to retire within the decade. One of the most promising areas is sometimes called seniorisation: helping a junior employee close the gap with a 20-year veteran far faster than usual. So how to use AI to address this? First, by ensuring that the operational knowledge of senior workers is adequately captured in procedures, practices and understanding of risk. Second, by connecting and contextualising information across various sources. Third, by ensuring that AI can effectively prepare, check, route, generate, recommend and prioritise work.
Octave
www.octave.com
automationmagazine.co.uk
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40
