By John Keilman | Photography and video by Angela Owens/WSJ
There's an iron law in the paint business: The faster it dries, the worse it looks.
But then AI took a crack at it.
Scientists at paint manufacturer PPG built a database containing the properties of all of their products, overlaid with the laws of chemistry. Then, about a year ago, they asked the system to create something new -- a fast-drying clear coat that body shops could apply after repainting a car.
Within minutes, the system suggested a combination of chemicals PPG's scientists hadn't thought of. Real-world testing proved that it worked, and last spring the clear coat, which cuts drying time by more than half, went on sale. It was PPG's first AI-assisted new product. Dozens more are in the pipeline.
"What's really exciting in this field is when the model picks formulations that you intuitively would not have," said Daniel Connor, the company's global technical director for automotive refinish coatings.
Manufacturers say that AI, known for creating instant term papers and pixel-perfect fake videos, is fundamentally changing how new products are created. It is letting companies speed-run a process that can often be a deliberative slog relying on tried-and-true approaches.
AI tools have helped Procter & Gamble create new scents for body washes, laundry beads and home fragrances. They have allowed Mars to design a thinner-walled bottle for its Extra brand chewing gum that reduced development time by 40% and saved 246 tons of plastic. And they have assisted 3M in coming up with a sanding disc that optimizes dust collection and grinding performance.
John Banovetz, 3M's chief technology officer, said AI is playing the role of an additional colleague.
"When I was in the lab, I might talk to three different experts about something," he said. "AI would just be the fourth expert I'd talk to."
'Impossible for a human'
Pittsburgh-based PPG has tight connections with Carnegie Mellon University, a national center of AI research. About four years ago, rumblings of the coming technology persuaded PPG to set up an internal team dedicated to exploring it, said Chief Technology Officer David Bem.
The product-development system PPG designed isn't a large language model that is subject to hallucinations, the make-believe answers that can plague chatbots. It relies on what is known as deterministic AI, in which outputs must conform to the laws of science.
The team spent several years making "digital twins" of PPG's products, mirroring their formulas and attributes, and then embedded algorithms reflecting chemistry principles. That gave the system the ability to create near-limitless blends virtually and to predict how they would perform, though any suggestion still had to be vetted in the lab.
PPG has a large automotive business, and in that industry, paint is applied in a "stack" of several coatings, each of which has its own properties. A single coating can have upward of 25 ingredients.
"It is impossible for a human to search every possible combination," said PPG technical manager Jun Deng. "So what AI or machine learning is doing is helping us to see this vast amount of potential candidates."
The company's test case was a protective clear coat, the final layer in a car's repainting. Its drying time can be a pinch point in a body shop's operations, limiting how many cars it can turn over in a day. The question for AI was how to achieve a flawless gloss or matte finish as quickly as possible.
PPG said its system found an unexpected chemical response that cut the post-spray drying time from 30 minutes to 5 when heated at 140 degrees. Air drying took less than an hour compared with two hours for competitors' products, PPG said.
The coating, dubbed Deltron Premium Glamour Speed Clearcoat, went on sale in March. PPG declined to specify sales but said it is performing well compared with similar products.
Andy Powell, owner of Andy's Auto Body in Vandalia, Ill., said the coating's fast-drying properties have worked as advertised since he added it to his shop's rotation.
"With this clear, we can throw in an extra job here and there, and it really helps us out," Powell said.
AI toolmakers
Other companies have brought in specialists like Citrine Informatics, which makes AI tools for chemical and material manufacturers. Greg Mulholland, Citrine's chief executive, said some companies are using AI to improve several attributes of a product simultaneously.
"The reality is, no one cares about a material that's just lighter or a material that's just stronger," he said. "You care about lighter and stronger and a dozen other things. And so AI is very good at keeping all of those in its head, so to speak."
One of Citrine's clients, specialty chemical maker Stepan, said the platform has allowed it to slash the development time of some new products from weeks to days. TerraSafe, a North Carolina-based startup that specializes in nonplastic food packaging, used it on a side project -- creating laundry detergent sheets that dissolve fully in the wash without sacrificing cleaning power.
Though the platform helped TerraSafe devise promising candidates, the company decided to end the effort to save money and focus on its main product line. CEO Julie Willoughby said she would be game to try it again once the company, which is still in the fundraising stage, is better positioned to turn digital ideas into physical prototypes.
"I still believe there's a place for generative AI in material discovery and design," she said.
Write to John Keilman at john.keilman@wsj.com
(END) Dow Jones Newswires
January 24, 2026 12:00 ET (17:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments