Ready to learn more?
Request a demo to learn how your team can accelerate adhesives R&D
Blog
Adhesives in Action at ASC: Reducing Real-World Development Timelines by 30%
No items found.
Albert Invent co-presented alongside Henkel at ASC 2025 to share real-world stories of accelerated innovation.
Albert Invent recently attended the Adhesive and Sealant Council’s (ASC) 2025 Annual Convention and Expo in Jacksonville, Florida. Zac Brown, Solutions Engineer at Albert, presented alongside Tim Champagne, Product Development Manager at Henkel, during the session Smart Tools – Artificial Intelligence, Machine Learning, and Data Management. The presentation highlighted examples of how data-driven R&D is accelerating innovation in real-world applications.
A connected, chemically intelligent platform
Zac began the presentation with a demonstration of how data-driven R&D can accelerate adhesives development. Using 21 epoxy adhesive formulations sourced from two journal articles and a patent, Zac entered formulations and glass transition temperatures (Tg) into Albert Breakthrough, our suite of AI/ML tools trained on over 15 million molecules.
Breakthrough not only developed a model, but it returned information about how well the model is performing and what features impact Tg the most.
“You’ll notice that the features returned by the model are molecular descriptors such as hydrogen bond donor count,” emphasized Zac. “These insights originate from the CAS information assigned to the inventory items used in the formulations and it allows us to make intelligent predictions across raw materials.”
In other words, in a connected, chemically intelligent platform, inventory items that are rich in chemical information are linked to formulations, which are in turn linked to performance properties. Zac demonstrated the power of this interconnectivity by removing a target formulation and four others with the closest performance. After training new models on the remaining data, Breakthrough successfully predicted which of the removed formulations had a Tg target of 150°C on the first or second guess 85% of the time – highlighting the effectiveness of a small but smart data set.
Faster, smarter development in the real world
In practice, this translates into faster product development timelines.
Tim shared a case study illustrating how AI-driven formulation decreased semiconductor product development timelines at Henkel by 30%.
In this example, Henkel was developing underfill technology, an encapsulant used in semiconductor packaging applications to fill the gap between the chip and substrate to protect delicate components. There was a challenging trade-off: the team needed to develop a one-component epoxy with high filler loading while still maintaining flow.
“Because we were able to upload our historical data into Albert, we could search more efficiently and start from a point that was much closer,” Tim noted. This head start had cascading effects, streamlining every stage of development and even enabling Henkel to complete their scale-up during the validation stage in two different locations.
Zac followed this story with a second real-world example where an Albert customer wanted to maintain a product’s weathering performance while reducing the ingredient correlated with long-term stability, such as an antioxidant or UV stabilizer. The machine learning model not only rapidly found a solution to this technical challenge, but it did so by exploring new materials.
“Because we’ve trained the model to use molecular descriptors, we can incorporate raw materials we’ve never used before,” said Zac. “As long as the CAS information is there, we can get intelligent predictions on a formulation’s performance, even on the first try.”
These real-world wins not only highlight reduced product development timelines, but they also demonstrate how an AI-powered platform can help scientists break performance barriers and explore new frontiers.
The future of adhesives R&D
The adhesives industry is undergoing a fundamental shift, moving away from traditional methods towards data-driven R&D. This approach connects data across raw materials, formulations, and test results in a way that enables scientists to make better predictions faster – ultimately allowing them to provide their customers with higher quality products on a shorter timeline.
If we didn’t get the chance to connect at ASC 2025, Request a Demo to learn how Albert can accelerate your adhesives R&D.
Albert Invent recently attended the Adhesive and Sealant Council’s (ASC) 2025 Annual Convention and Expo in Jacksonville, Florida. Zac Brown, Solutions Engineer at Albert, presented alongside Tim Champagne, Product Development Manager at Henkel, during the session Smart Tools – Artificial Intelligence, Machine Learning, and Data Management. The presentation highlighted examples of how data-driven R&D is accelerating innovation in real-world applications.
A connected, chemically intelligent platform
Zac began the presentation with a demonstration of how data-driven R&D can accelerate adhesives development. Using 21 epoxy adhesive formulations sourced from two journal articles and a patent, Zac entered formulations and glass transition temperatures (Tg) into Albert Breakthrough, our suite of AI/ML tools trained on over 15 million molecules.
Breakthrough not only developed a model, but it returned information about how well the model is performing and what features impact Tg the most.
“You’ll notice that the features returned by the model are molecular descriptors such as hydrogen bond donor count,” emphasized Zac. “These insights originate from the CAS information assigned to the inventory items used in the formulations and it allows us to make intelligent predictions across raw materials.”
In other words, in a connected, chemically intelligent platform, inventory items that are rich in chemical information are linked to formulations, which are in turn linked to performance properties. Zac demonstrated the power of this interconnectivity by removing a target formulation and four others with the closest performance. After training new models on the remaining data, Breakthrough successfully predicted which of the removed formulations had a Tg target of 150°C on the first or second guess 85% of the time – highlighting the effectiveness of a small but smart data set.
Faster, smarter development in the real world
In practice, this translates into faster product development timelines.
Tim shared a case study illustrating how AI-driven formulation decreased semiconductor product development timelines at Henkel by 30%.
In this example, Henkel was developing underfill technology, an encapsulant used in semiconductor packaging applications to fill the gap between the chip and substrate to protect delicate components. There was a challenging trade-off: the team needed to develop a one-component epoxy with high filler loading while still maintaining flow.
“Because we were able to upload our historical data into Albert, we could search more efficiently and start from a point that was much closer,” Tim noted. This head start had cascading effects, streamlining every stage of development and even enabling Henkel to complete their scale-up during the validation stage in two different locations.
Zac followed this story with a second real-world example where an Albert customer wanted to maintain a product’s weathering performance while reducing the ingredient correlated with long-term stability, such as an antioxidant or UV stabilizer. The machine learning model not only rapidly found a solution to this technical challenge, but it did so by exploring new materials.
“Because we’ve trained the model to use molecular descriptors, we can incorporate raw materials we’ve never used before,” said Zac. “As long as the CAS information is there, we can get intelligent predictions on a formulation’s performance, even on the first try.”
These real-world wins not only highlight reduced product development timelines, but they also demonstrate how an AI-powered platform can help scientists break performance barriers and explore new frontiers.
The future of adhesives R&D
The adhesives industry is undergoing a fundamental shift, moving away from traditional methods towards data-driven R&D. This approach connects data across raw materials, formulations, and test results in a way that enables scientists to make better predictions faster – ultimately allowing them to provide their customers with higher quality products on a shorter timeline.
If we didn’t get the chance to connect at ASC 2025, Request a Demo to learn how Albert can accelerate your adhesives R&D.