The team at Tenor came together to see if they could simplify the process of documenting physician consultations with the power of Artificial Intelligence. One of their challenges was the inaccuracy of Natural Language Processing, and how to empower physicians to moderate the output without increasing their workload or needing to invest in new devices.
- Devise simple workflow for capturing/editing data
- Craft interaction design models for various types of anticipated errors
- Design models for both listening and moderating
- Create a design system for low-fidelity prototyping
The vision for Tenor was both realistic in addressing a real-world need, yet ambitious in its challenge to improve the consultation process of physicians without disrupting their workflow. Mark truly understood the needs of our audience, and helped us realize our vision with a usable, meaningful, and thorough app experience that we could test with our physician partners, on a very abbreviated timeline.
Mike Sousa, CEO
My initial course of action was to conduct research with the aim of verifying Tenor’s theories about how physicians might adapt to a modified workflow, which would begin by researching the fundamental consultation processes of physicians. I also wanted to get a consensus (if possible) about potential trust/privacy concerns of physicians and patients, primarily with having an active microphone present during a patient consultation, which would require end-user interviews. Finally, I wanted to better understand the common types of Automatic Speech Recognition and Natural Language Processing errors identified by Tenor’s engineer so I could craft a contingency design plan for these use case scenarios.
Scope of Work
- Generative research
- Use Case Scenarios
- End-user interviews
- Journey map
- Task flows
- Usability testing
- UI prototyping
Working closely with the Tenor’s engineer, founder, and physician advisor, we were able to craft a complete journey through a typical consultation process, while accounting for some secondary variances. The journey map enabled us to build a holistic experience for the three primary scenarios of interaction between physicians and the system and build a respective prototype to share with investors and test with physicians, as a polished proof-of-concept.
- Usability-test task success rate
- Common voice errors rectified