AI‑powered formulation tools for beauty and personal care R&D

We use data to build digital tools that formulate better products, faster.

Our customers exist across the BPC value chain‑from ingredient suppliers and contract manufacturers to brands and retailers.

Predictive analytics

Predict R&D outcomes digitally

Shorten and streamline the product development process.

Simulate product outcomes before performing lengthy testing procedures (e.g., stability, compatibility, etc.)

Evaluate endpoints such as efficacy, texture, physiochemical characteristics, etc. prior to any benchwork or testing

Virtual formulation

Generate formulas with the click of a button.

Innovate faster and better

Leverage AI‑based insights to expand beyond the limitations of human heuristics and push the boundaries of formulation innovation

Optimize for several product attributes at once, automating up to 80% of the formulation process

Quickly reformulate existing products or find ingredient substitutions to match new requirements

Data management

Create a single source of R&D truth.

Leverage valuable latent data regardless of your company's current technological infrastructure

1. Future‑Proof Organizational Knowledge

Digitize historical paper / analog R&D data into a ready‑to‑use format

Clean, structure, and consolidate all organizational R&D data into a single source for the first time

Prepare to extract valuable insights through data analytics, data science, and AI / ML applications

Technology

The Future of Formulation.

Using deep learning frameworks‑akin to those used in de novo drug design‑we're building the world’s most advanced formulatory artificial intelligence.

Our patent‑pending technology is the first commercial use of general adversarial networks (GANs) and recurrent neural networks (RNNs) at a formulation level.

Scalable Performance

Deep neural networks are the first class of algorithms for which more data = better performance

Self‑Taught

Deep learning breaks down complex concepts into simpler ones, removing the need for human‑led feature extraction (i.e., input of rules, etc.)

How we work

We believe in an open innovation model for formulation in beauty, personal care, and beyond. We work with our customers in two ways:

Consortium

Customers contribute data to a pool‑resulting in superior formulation tools, lower R&D costs, and a competitive advantage over non‑consortium peers

Customer 1

Customer 2

Customer 3

Exclusive

Customers rely on their internally available data and expertise to power Potion's formulation tools

Contact

hello@potion.ai

Careers linkedin