Build trust in AI-enabled SaMD with clinician-driven design

Our client set out to design an AI-enabled software as a medical device (SaMD) to assess pregnancy risk and predict adverse outcomes using patient data from primary and secondary care settings. To succeed, they needed an understanding of clinicians’ goals, workflows, and behaviors, and to ensure the tool would be trusted and adopted in maternal care.

We led the full research and design process in partnership with a global design agency. Using an agile, iterative approach, we ran three design cycles that included in-depth interviews, two clinical advisory boards, and discovery workshops with designers, researchers, a midwife, and the client team to ground the work in real clinical practice. Through workflow analysis and user interviews, we uncovered pain points in current risk assessment practices and used that insight to iteratively design and test mid-fidelity prototypes. Each round of feedback helped refine the tool to better align with clinicians’ mental models and day-to-day routines.

By the third cycle of research, clinicians expressed confidence that the tool could reduce maternal risks and integrate smoothly into care delivery. Their feedback shaped a minimum viable product (MVP) that addressed both functional needs and trust in AI recommendations.

The insight gave our client the confidence to move forward. They developed an MVP and advanced to the next stage of SaMD development, backed by clinician buy-in and real-world insight into adoption potential.