With Robots in Control, Chemical Makers Fight for Their Formulas

(Bloomberg) -- Four decades ago as a young engineer working for Imperial Chemical Industries Ltd., Patrick Thomas helped install one of the sector’s first digital computer systems: a mahogany-encased machine now on display in a science museum.

Computing has since transformed the industry into a $4 trillion business supplying ingredients for all types of manufactured goods, ranging from shampoo to paint to mattresses. Yet for Thomas, it’s also emerging as a battleground over ownership of valuable formulations.

“There’s no sort of regulatory framework and there’s no shortage of lawyers trying to get involved,” said the executive, who headed German chemicals maker Covestro AG for more than a decade before becoming chairman of Johnson Matthey Plc on June 1. “Who owns the data in the chain?”

The warning from the top industry veteran serves to highlight a debate raging across the corporate world over data, privacy and ownership rights, which has particular relevance to chemical makers because of their growing reliance on factory automation. Industrial robots are stepping into the shoes of trusted employees, absorbing huge quantities of sensitive data from both manufacturers and customers, while learning and modifying key formulations along the way.

Capture Value

Suppliers have captured all the value created by artificial intelligence, according to Vijay Sarathy, head of chemicals for North America at consultancy firm Accenture. Customers contributing data to the manufacturing chain now feel they are owed something for their input.

“One thing that won’t work is someone thinking that they can keep all the value created to themselves,” he said. The ownership lines are especially blurred for the chemicals sector because of its highly interactive relationship with customers, reliance on broad supply chains and input from software companies helping to develop automation tools and products.

The first automated machine controls for chemicals began in the 1950s, followed by development of software from companies including SAP SE and Oracle Corp. to run whole plants. The latest offerings -- part of a broader move to digitalization known as Industry 4.0 -- embodies self-learning robots designed for the autonomous plants of the future.

Tighten Rules

Although Europe this month tightened the rules on data protection for individuals, chemical manufacturing is still an unregulated wild west, Thomas said in the interview. It’s getting more difficult to protect a company’s know-how, while the lack of rules on data ownership is slowing advances and spawning defensive moves by companies seeking to protect their innovations.

Companies are experimenting “to see what works,” he said. In the case of Covestro, the German maker of plastics and coatings spun out of Bayer AG, customers are offered a kind of digital dashboard to adjust formulations to alter glossiness, color-fastness and product durability, but the precise data recipe is kept in a detached cloud owned by the company.

The arrival of artificial intelligence onto factory floors is timely as experienced workers reach retirement age, executives said. Self-learning software can go some way in filling the gap left by long-standing workers who could often adjust machines or formulas almost instinctively.

“I’ve worked at a pharmaceutical site like that, where your packaging operators would come in and listen to a filling line and know it’s not running right,” said Richard Henderson, automation manager at paintmaker Akzo Nobel NV’s new state-of-the-art plant in Ashington, England. “We want that built into the process and basic controls of the system, so it’s not down to the individuals to make those judgments.”

To contact the reporter on this story: Andrew Noël in London at anoel@bloomberg.net

To contact the editors responsible for this story: Anthony Palazzo at apalazzo@bloomberg.net, Tara Patel

©2018 Bloomberg L.P.