Every Peak Has a Story
What we learned presenting with Coca-Cola at AOAC 2025
At AOAC INTERNATIONAL 2025, Expert Intelligence joined The Coca-Cola Company to present a real-world example of how AI can be applied inside a regulated analytical workflow.
The session, Natural Oil and Wine Authentication Using Deep Generative AI, explored how AI can help labs see deeper into GC-FID and LC-MS/MS data, not by replacing expertise, but by making expert judgment scalable, consistent, and auditable.
In many labs, the challenge is not generating data.
It is interpreting it reliably, repeatedly, and under regulatory constraint.
Every chromatogram, every spectrum, every peak carries context that experienced analysts understand but traditional automation struggles to capture.
That is the gap we set out to address together.
Working with the Coca-Cola team, we applied Limited Sample Model (LSM), our AI architecture designed for sparse, high-value datasets, to support authentication workflows where accuracy, traceability, and defensibility matter as much as speed.
The goal was not to build another black-box model.
It was to create a system that could learn from expert decisions, apply that judgment consistently across large datasets, and keep scientists in the loop when uncertainty required review.
What made this collaboration meaningful was not the technology alone. It was the fact that the work happened inside a real laboratory environment, with real constraints, real data, and real quality requirements.
This is where AI either works, or it doesn’t.
Sessions like this reflect a shift happening across the industry.
AI adoption in regulated labs is moving from experimentation to infrastructure.
From proof-of-concept to production.
From curiosity to necessity.
We are grateful to the AOAC community and to our colleagues at The Coca-Cola Company for the opportunity to share this work, and to continue the conversation about what trustworthy, governed AI should look like in modern analytical labs.
#AOAC2025 #AnalyticalChemistry #FoodSafety #LabAutomation #AIinScience #TrustedAI #MassSpectrometry #QualityControl #DataAnalytics
