In process development, the bottleneck is rarely ideas — it is the number of bioreactor runs you can afford, and how quickly you can learn from each one. Every campaign costs time and money, and without a structured way to learn across experiments, you end up repeating work or missing parameter interactions that only become visible when the data is properly organised.
What Hercule brought us is a methodology as much as a platform. Our development work runs on an Ambr250 — up to 24 parallel bioreactor conditions per campaign. Without a common data model, the output of an Ambr250 campaign is a large, unwieldy timeseries dump across 24 reactors. With Hercule, every condition is mapped to the same ISA S88 process hierarchy from the start: same parameter names, same phase structure, directly comparable across all conditions from the moment data is captured. Each campaign produces a dataset that is immediately analysis-ready — no harmonisation step before you can ask a question.
That structured foundation is what allows you to identify parameter interactions and operating ranges faster, carry knowledge forward from one campaign to the next, and move into CPV without re-mapping when the process reaches the clinic or manufacturing.
On our first project using the Hercule methodology, we went from 45 weeks of development to 28 weeks — a 17-week reduction. For client programmes where timelines are contractual and Ambr250 campaigns are expensive to repeat, that difference is substantial. Review collected by and hosted on G2.com.
The methodology requires investment upfront: structuring the process in ISA S88 before the first run, defining which parameters will be captured and how. That discipline is what makes everything downstream work, but it requires the development team to think carefully before running experiments rather than capturing data opportunistically. If your team is accustomed to flexible, ad hoc data collection, the initial adjustment takes some time. Review collected by and hosted on G2.com.
