We scan new podcasts and send you the top 5 insights daily.
Christian Witz notes that biotech scientists frequently treat CFD as an intimidating tool requiring a dedicated physics or mathematics PhD. However, the real obstacle is that conventional CFD user interfaces have not been tailored to bioprocess workflows. Just as modern computer use does not require an IT technician, tailored interfaces allow routine process development scientists to employ CFD daily like Excel.
Technical interfaces like drag-and-drop workflow builders are immediately rejected and delegated by senior business leaders. To achieve executive buy-in and direct engagement with AI process tools, the interface must be presented in a familiar format: a plain English document that they can read and edit.
Powerful AI models for biology exist, but the industry lacks a breakthrough user interface—a "ChatGPT for science"—that makes them accessible, trustworthy, and integrated into wet lab scientists' workflows. This adoption and translation problem is the biggest hurdle, not the raw capability of the AI models themselves.
Because CFD operates strictly on physical parameters rather than biological mechanisms, a vessel needs to be fully characterized only once across 100 to 200 operational points. When switching to a new molecule, media, or cell line, the scale-up software only needs updated fluid properties such as viscosity, density, surface tension, and salt concentration, eliminating the need to re-run simulations from scratch.
Early CFD adoption in biotech faltered because limited compute power forced engineers to use geometric simplifications, such as simulating a single periodic cake slice instead of the entire reactor cylinder. Bioreactors contain massive three-dimensional wrapping flows that cannot be captured in periodic wedge models. These simplifications produced inaccurate results, leading teams to experience poor outcomes and abandon simulation for years.
A technology's success isn't just about high-level metrics like facility output. If the product is inconvenient or frustrating for the end-user—the scientist at the bench—it will face strong resistance and likely fail, regardless of its theoretical benefits. User experience is paramount.
Characterizing a bioreactor requires careful boundary setting because every additional variable exponentially increases the required simulation count. Adding wide continuum ranges for parameters like working volume or fluid viscosities creates an enormous operational space. Teams must strategically limit variables to discrete, sensible operating increments rather than attempting to model extreme spectrums from empty vessels to honey-like viscosities.
The most significant breakthroughs will no longer come from traditional wet lab experiments alone. Instead, progress will be driven by the smarter application of AI and simulations, with future bioreactors being as much digital as they are physical.
The future of bioprocess development involves using AI on high-throughput data for predictive modeling. This, combined with in silico simulations (digital twins), will allow scientists to understand underlying biological mechanisms, not just identify optimal conditions, dramatically accelerating optimization.
Scientists won't adopt automation if they have to code or use clunky visual programmers. The breakthrough is using AI models to translate natural language protocols into robot commands. This removes the primary usability barrier and prevents common user errors, enabling adoption.
Despite powerful capabilities, AI tools remain largely inaccessible to non-technical users due to complex interfaces and frustrating setup processes. The industry's focus on technical prowess over user-centric design is the primary obstacle to widespread adoption in business workflows.