For most of history, medicines came from plants. Aspirin descends from willow bark, morphine from the poppy, and paclitaxel, a cornerstone of cancer treatment, from the Pacific yew. Yet for decades the industry moved away from nature. Plants had not run out of answers; finding those answers was simply too slow and too expensive. Discovery meant screening thousands of extracts, spending years to isolate a single compound, and investing millions before knowing whether it worked.
That equation is changing fast with AI.
Reading the chemistry of a plant
Modern mass spectrometry captures thousands of molecular signals from a single extract. Open, community-curated libraries such as GNPS let researchers compare those signals with data shared by laboratories worldwide. Machine-learning models can now propose structures for compounds that have never been described. Work that once took years of isolation can now start as an annotated map of a plant's chemistry, built in weeks. AI helps to find new active molecules (APIs) in plants with accurate efficacy.
Predicting what a molecule does with AI, before the lab
In 2022, AlphaFold made more than 200 million predicted protein structures freely available to every researcher. Combined with AI target-prediction and docking tools, this lets scientists explore in silico how a compound may interact with a biological target. They can then rank candidates before running a single bioassay. AI-led discovery has already reached patients: in 2025, rentosertib, a drug whose target and molecule were both identified with AI, reported positive Phase IIa results in Nature Medicine. Capabilities once reserved for the largest pharmaceutical groups are now broadly accessible.
Nature as the next AI-based chemical library
Natural compounds have been shaped by millions of years of evolution to interact with living systems. They offer a structural diversity that synthetic libraries struggle to match. AI does not replace experiments; it makes them targeted, faster and far less costly. Now that these tools are broadly available, the question is no longer whether natural compounds can compete with synthetic molecules, but which ones, and for which uses. In health, nutrition and cosmetics, a future where natural actives take the place of synthetic molecules is moving within reach.
One condition remains for this future to reach products: the same plant must deliver the same chemistry, batch after batch. That is why, at BotaniX, we grow our plants in controlled conditions, so that what AI discovers can be reproduced.
