Swiss AI startup Devia aims to automate quality management in the pharmaceutical industry

Devia, based in Zurich, is making its first public appearance following a year-long development and pilot phase. With a specialized AI platform, the startup aims to accelerate the investigation of quality deviations in the pharmaceutical industry and detect recurring errors early on. According to the company, initial pilot projects show that the investigative effort can be reduced by up to 50 percent.

Quality deviations are among the most resource-intensive challenges in pharmaceutical production. When manufacturing processes do not proceed as planned, companies must identify the causes, assess potential impacts, and document appropriate corrective actions. These investigations are not only time-consuming but are also subject to strict regulatory requirements. According to the company, previous customer pilot projects have halved the time required for investigations.

Regulated AI as an Opportunity for Pharmaceutical Companies

Unlike many general-purpose AI applications, Devia was developed specifically for regulated production environments. The platform is designed to meet the requirements of Good Practices (GxP), which are intended, among other things, to ensure the quality and traceability of pharmaceutical processes. Every AI recommendation is documented. Qualified employees review and approve each step. This ensures transparency regarding which recommendations the system made and which decisions were actually implemented.

Devia can also be integrated into existing quality management systems such as TrackWise, Veeva Vault QMS, and MasterControl. This means companies do not have to completely replace their existing systems. The target audience includes large biopharmaceutical companies, contract manufacturers, and contract research organizations (CROs).

Regulatory developments also play a role in the business model. With the European AI Act, requirements for certain AI applications are increasing, particularly with regard to transparency, documentation, and risk management. For Devia, this presents a market opportunity. The platform is designed not only to support companies in automation but also to improve the quality and consistency of their studies. It remains to be seen whether the results of the pilot projects will be confirmed in widespread use. However, the approach demonstrates how specialized AI agents are increasingly penetrating highly regulated industrial sectors and can take on specific operational tasks there.