At Potato AI, we build tools to distill scientific content from literature and help researchers plan and run experiments. We help researchers identify new hypotheses and plan step-by-step wet lab protocols and computational workflows using plain language. Our customers are biotechs, pharma, and academic labs. We are partnering closely with scientific publishers to expand our access to other scientific specialties like chemistry and materials science.
What is the problem?
Scientists are overwhelmed with the amount of new scientific information in the literature. Meanwhile, the majority of published papers can't be replicated, often because individual papers don't contain enough information. As a result, scientists spend too much time scouring the literature for experimental details. We believe this inefficiency and the lack of experimental reproducibility is holding back science.
What is their solution?
We've built proprietary AI tools to help with various parts of the scientific workflow. We've implemented RAG (retrieval augmented generation) to search for relevant literature and ensure information is accurate and can be cited. We're also building proprietary data extraction technology to structure unstructured scientific information, so we can compare experimental details across multiple protocols.