Lab scientists built an AI system to speed up ingredient development. It cut five years of research to one hour.
Applied Laboratory Technologies, founded late last year by University of Guelph scientist Dr. Paul Spagnuolo and IT specialist Brian Johnston, created AMBROSIA-an AI system designed to accelerate the ingredient development process by automating molecular target identification, pathway analysis, and mechanistic hypothesis building.
The system emerged from frustration. Spagnuolo spent years developing avocatin B from avocado for metabolic health through traditional lab methods-a process he described as "finding a light switch in the dark," probing multiple directions until discovering something that works.
When Spagnuolo discussed these bottlenecks with Johnston, the pair saw an opportunity to computerize the workflow. "It was kind of like trying to figure out how to virtualize Paul's brain," Johnston said. He built a proof of concept by translating the established research blueprint into an autonomous computational engine.
Testing the system on Spagnuolo's five-year avocatin B research produced the same results in one hour.
What the system does and doesn't do
AMBROSIA works at the molecular level to characterize plant extracts and identify new markets for existing ingredients. It won't work for probiotics, but it can analyze postbiotic metabolites and combinations of molecules in future versions.
Results still require lab confirmation. The system operates within defined scientific guardrails, not as a replacement for experimental validation.
Early adoption from ingredient suppliers
Finzelberg, a German-based ingredient supplier, became the first client before the company was officially incorporated. The company used AMBROSIA to analyze Menofelis, an extract from Siberian rhubarb for menopause support, identifying multi-target profiles across hormone signaling, inflammation, and detoxification.
Dr. René Roth-Ehrang, member of Finzelberg's management board for quality and development, said the system "accelerates the development of effective botanical solutions, ensuring that Finzelberg's products set new standards for natural health and scientific credibility."
Spagnuolo presented the concept to dsm-firmenich and Nestlé at their Switzerland headquarters following an invitation to present at the International Conference on Food Supplements in November 2025.
Repurposing ingredients as a market opportunity
One feature attracting client interest is the ability to discover new applications for established ingredients. Spagnuolo, who completed postdoctoral work in drug repurposing, sees parallel potential here: AMBROSIA can identify where an ingredient might serve a different market than originally intended.
This capability addresses a real product development challenge-finding new revenue streams from existing ingredient portfolios without starting from scratch.
What sets it apart
Johnston emphasized the company's approach: "Our differentiating factor is we built ours because a PhD had a problem." The system was built on methods that worked in actual labs, not designed in isolation from scientific practice.
"The tagline on the website is 'Where laboratory science meets computational power,'" Johnston said. "We put laboratory science first."
For product development teams evaluating AI tools, the distinction matters. This system was built to solve a specific workflow problem in ingredient development, not adapted from a general-purpose platform.
Your membership also unlocks:
