The future of biological experiments has to involve the routine use of automation and artificial intelligence (AI). As an industry, we need to use and understand ever more complex biological systems in our efforts to conduct experiments that are as relevant to the clinic as possible. The only way we will get a handle on these systems efficiently and effectively is by making the most of powerful hardware and software. We’re already well on the way to mastering automation, but what will it take to truly bring AI into our experimentation? With AI, we could run and understand our experiments in new ways: knowing the best experiment to run next, knowing the best way to improve and optimise our experiments, and knowing how to make the most out of the data produced. But there are significant barriers in the way. The biggest is the lack of properly curated high-quality data. This means both the data from experiments, and the data about experiments (metadata). There are limits on what we can record in any given experiment for multiple reasons. Primary among them is often the constraint of what a single scientist can record. Here we present how a detailed digital experiment model (DEM) can be programmatically generated, used to drive automation, and provide a highly detailed, multi-layered record of what has been carried out. In this way, a DEM can provide the highly structured link that is needed between complex, automated experiments and AI, which will enable the routine use of AI to drive rapid, conclusive insights from experiments.