RoslinCT has secured a Scottish Enterprise research and development grant of up to £7.25m for a multi-year programme to expand digital and artificial intelligence capability across its cell therapy manufacturing operation in Scotland. The project is intended to connect manufacturing, analytical and process-development data more effectively as the contract development and manufacturing organisation supports therapies moving from clinical work towards repeatable commercial production.
Cell therapy creates a particularly demanding manufacturing record because the product is inseparable from the process used to make it. Material identity, analytical results, equipment conditions, operator actions and batch history all contribute to the evidence required to understand whether a manufacturing run remains within control, and those records become harder to review when information is divided between systems or transferred manually. RoslinCT’s programme is therefore centred on the data architecture beneath the production process rather than simply adding an AI tool to an existing workflow.
The company says the work will extend across process development, analytical development, GMP manufacturing and commercial readiness, linking stages that generate different types of evidence but ultimately describe the same product. Better continuity between those datasets can make technology transfer more systematic because engineers can trace how parameters established during development behave once a process is operating in a regulated manufacturing suite, rather than reconstructing that relationship from separate records.
Artificial intelligence enters the programme after that data problem, not before it. RoslinCT has not disclosed individual models or autonomous production applications, and the grant should not be read as a commitment to hand manufacturing decisions to an algorithm. Analytical tools become useful only when equipment, batches, materials and test results are identified consistently enough for patterns to be compared, making integration and data quality prerequisites for any later AI layer.
Commercial cell therapy production increases that pressure because the tolerance for informal workarounds declines as volumes and batch numbers rise. A clinical programme can rely on specialist staff manually reconciling information across a relatively small number of runs, whereas a commercial operation has to repeat the same controls efficiently while preserving traceability, deviation management and review. Reducing duplicated data entry or improving access to process histories can therefore remove administrative friction without weakening the GMP requirements governing release.
RoslinCT operates 22 purpose-built cGMP cell therapy suites across Edinburgh and Hopkinton, Massachusetts, although the Scottish Enterprise award is directed at the company’s operations in Scotland. Its technical work covers autologous and allogeneic therapies, induced pluripotent stem cells, gene editing and differentiation, which means one digital architecture has to accommodate several manufacturing models rather than forcing them into an identical process.
That variety also makes comparison useful. When a manufacturer can connect process conditions with analytical results across repeated runs, engineers have a stronger basis for identifying drift, recurring deviations or steps that contribute disproportionately to variability. A correlation inside the data does not automatically justify a process change, particularly in regulated manufacturing, but it can tell development and production teams where controlled investigation is most worthwhile.
The same infrastructure can support manufacturing scale without claiming that software creates physical capacity by itself. RoslinCT has not published a quantified increase in suite throughput, batch numbers or release speed, and the grant value cannot be translated directly into additional manufacturing output. Digital systems can reduce review time and manual handling, but cell culture duration, equipment availability, cleanroom capacity, quality controls and staffing continue to set hard limits around production.
Scottish Enterprise has capped its contribution at £7.25m, while RoslinCT has not disclosed the full programme cost or a system-by-system implementation timetable. That leaves the project’s financial scale only partly visible, although the multi-year structure indicates that the company is treating digital integration as an operating programme rather than a short software installation followed by handover.
The useful test will come once the new data environment is embedded in routine GMP work. If engineers can move from development into production with fewer disconnected records, review batch performance without rebuilding datasets manually and investigate deviations against a more complete manufacturing history, the programme will have altered how the plant works; if the systems merely collect more information without improving those decisions, £7.25m of public support will have produced a more expensive archive rather than a more capable factory.
