Artificial intelligence (AI) is top of mind for many scientists in the chemical industry, but there remains a gap between interest and implementation. Nick Gripp, Senior Chemist at Applied Molecules, bridges that gap: he integrates AI into his day-to-day lab work, beginning nearly every experiment with AI-driven predictions and constantly feeding new data back into his models. Nick represents what many in the industry believe the chemist of the future will look like – the difference is, he’s doing it today.
Applied Molecules develops coatings, adhesives, sealants, and elastomers for a wide range of industrial applications. Their research team is lean, yet they’ve punched well above their weight by embracing digital tools early on. After building up their foundation of structured data, Applied Molecules began layering Breakthrough, Albert’s suite of domain-specific AI/ML tools, into their workflows in 2024 – and quickly recognized its potential to institutionalize knowledge and accelerate product development.
“That first time we used Breakthrough for development was eye-opening,” Nick recalled. “Since then, we’ve used it on dozens of projects – it’s definitely become a coworker.”
Nick describes Breakthrough as a coworker because it doesn’t replace his role as a chemist – he guides it through iterative loops while defining design constraints, setting targets, and choosing which candidates to test throughout the process. With each new set of empirical data, he feeds information back into the model and restarts the loop, a process known as active learning.
“It’s particularly powerful when I’m entering a new space,” Nick said. “It helps me reach my targets faster.”
In one recent project, Nick used Breakthrough to simultaneously optimize a simple target property with a complex target property that is a function of multiple variables – in other words, a pair of properties that would not be trivial to optimize manually with intuition and a spreadsheet.
“Without Breakthrough, it would be shots in the dark,” said Nick.
Nick started by loading 10 historical formulations into Breakthrough that he felt would be most likely to hit the mark and setting up the model with constraints and targets. From there, Breakthrough generated 20 formulation candidates which were made, tested, and fed back into the model.
On the second iteration, Breakthrough suggested a range of candidates that seemed to cast a wider net, with some falling within spec and others falling further from the target – but after testing those and retraining the model with the new data, Breakthrough’s third iteration included multiple candidates that hit the simple target property exactly while satisfying the spec range for the complex property at the same time.
What stood out to Nick wasn’t just the speed and creativity, but also the effortlessness. “It’s such a stress reliever,” Nick said. “It’s not just how much more efficient it is – it’s how hands-off it can be. It frees up your brain for the creative work.”
Nick isn’t alone in his AI-driven workflows – at Applied Molecules, it’s part of how the organization operates. There’s an understanding across the team that everyone’s projects feed new data back into Breakthrough and that every result works towards sharpening the model’s accuracy.
“We’re starting to create a database over time, and it’s just making Breakthrough better and better,” Nick said.
It is even part of the team’s best practices to test every new raw material they bring in with a base formulation and feed the data into Breakthrough – a habit that not only compounds in value over time, but illustrates Applied Molecules’ dedication to AI-driven innovation.
Nick reflected that the team’s active use of Breakthrough makes him less frustrated about experiments that don’t go as planned. “In R&D, the reality is that there are a lot of failures,” Nick said. “But with Breakthrough, failures become valuable because they help train the model.”
By embedding AI into their everyday workflows, Applied Molecules is showing what the future of R&D already looks like: scientists who understand how to guide AI and work alongside it, integrating it into a modern version of the scientific method.
“I don’t feel like AI is replacing me,” Nick said. “It augments what I can do. Leaders across the industry are already adopting it, so becoming an AI expert is essential to stay ahead.”
AI may not be a magic wand, but it’s a powerful multiplier – especially for teams who invest in structured data and know how to interpret its insights. Scientists like Nick Gripp represent a new kind of chemist: one who uses data and domain knowledge together to invent faster, smarter, and more creatively than ever before.
.png)
.png)
.svg)