Operating Division: EvoPhase

From food and pharmaceuticals to aerospace, artificial intelligence could transform the efficiency of industrial processing. 76% of industrial manufacturing firms in one KPMG survey said their workforce is motivated to embrace cutting-edge technology.

“Many of the fields we operate in, like food, pharma, and chemical sectors, deal with particulate matter, and their systems have barely changed in decades,” says Professor Kit Windows-Yule, Chief Scientific Officer of EvoPhase. “One of our partners put it like this: in the past 50 years, the only thing that’s changed on most process equipment is the paint.” 

EvoPhase is applying an AI-led “evolutionary design” approach to industrial equipment. The team uses real world measurements to produce accurate simulations, and combines them with flexible geometry descriptions and predictive modelling to help industries redesign their process equipment to be more energy-efficient, productive, and sustainable.

At the core of EvoPhase’s technology are evolutionary algorithms that mimic the principles of natural selection. Combined with physical simulations, these algorithms can optimise the geometry of complex systems like bladed mixers, stirred tanks, or fluidised beds. Whether the system stores, stirs, blends, or moves granular materials, EvoPhase can identify how to improve it across equipment categories such as mills, dryers, roasters, and more. “With rising energy costs, the push for sustainability, and declining productivity, the idea that you can optimise a system just by changing its geometry is incredibly compelling.”

One of the company’s most fascinating technical innovations was the development of geometry kernels: mathematical representations that allow AI to describe and optimise virtually any shape or configuration. “We now have a single mathematical function that can describe any wind turbine or mixing blade ever built, and an infinite number that haven’t yet been,” says Professor Windows-Yule. “This means our system can generate new designs, simulate their performance, and recommend the most effective ones, all while allowing human engineers to inspect, understand, and refine the output.” This human-in-the-loop approach to AI has made EvoPhase promising for safety-critical, transparency-focused sectors like pharma.

EvoPhase was launched as an Operating Division (OD), a unique model that allows academics to test commercial ideas in a low-risk environment before spinning them out into independent companies. Team members also completed Enterprise’s Medici programme, which supports researchers in commercialising academic work.

The team behind EvoPhase has a deep understanding of particle and fluid dynamics, having been one of the first research groups in the UK to explore the intersection of AI and particle technologies or complex fluid dynamics. But Enterprise’s support has been crucial in translating this knowledge into a commercial and marketing context, including through guidance from John Cooke and Laura Bond, Entrepreneurs-in-Residence at the University. “While we’re probably among the more commercially engaged academics, we still come from an academic mindset,” says Professor Windows-Yule. “John and Laura bring a completely different perspective in business, investment, and sales, with knowledge of the different business models and the ways to deliver them.”

The team also credits Academic Commercial Services (ACS) with contracting and operational support, particularly when pursuing projects with major companies. “A process that might take six months through other channels can often be done in two weeks with ACS,” says Professor Windows-Yule. “For instance, they moved heaven and earth to secure us a contract with a well-known pharmaceutical company – I’ve never seen anything like it.” EvoPhase worked with the company to adapt a mill to simultaneously mix and coat small particles with a pharmaceutical ingredient, repurposing EvoPhase’s wind turbine geometry kernel to do so.

EvoPhase’s potential applications span an impressive range of industries. “If it has fluids that flow or particles that move, we can optimise it,” says Professor Windows-Yule. “We’re working with or in discussions with companies across aerospace, chemicals, pharmaceuticals, food and beverage, and others.” This has partly been built on his own academic and industry network, but CEO Dominik Werner is opening up new avenues. Successful projects have additionally generated publicity that has brought new clients.

The Birmingham Blade project was the world’s first AI-designed, geographically tailored urban wind turbine, in collaboration with local manufacturing company Kwikfab, who have since founded a spinout dedicated to producing wind turbines. Since the project’s reveal in 2024, interest in EvoPhase’s wind turbine solutions has exploded, from local authorities as well as industry. Even yacht designers have approached EvoPhase to explore turbine use on vessels.

As EvoPhase prepares to spin out, the team is taking a measured approach. “We’re holding the trigger for now, because the OD model is doing a lot of good and we don’t want to jump out sooner than we have to,” says Professor Windows-Yule. “But over the next 12 to 18 months, we’re putting the final tweaks on our business plan before approaching investors, and planning to roll out our software-as-a-service platform and test it with initial companies.”