Back

Company

Philips

Year

2024

EMR AI Assistant: Building Trust in Clinical AI

Context

Doctors were spending over half their workday on administrative tasks rather than patient care. The challenge wasn't just building AI-powered documentation for the Tasy EMR. It was earning clinicians' trust in a tool that would listen to consultations and auto-populate medical records. Usability testing with real physicians revealed immediate resistance: they didn't trust AI in clinical contexts.

My Role

As Product Designer for the Tasy EMR AI Assistant over two years, I was responsible for the voice assistant interface, AI design guidelines, and ambient documentation UX.

I led usability testing with physicians and designed the visual language that would differentiate AI-generated content from validated records.

Solution

One insight changed everything: doctors accepted AI as a copilot when validation stayed in their hands. We tested multiple visual approaches, starting with blue elements that disappeared against Philips' brand colors. The breakthrough came with the purple star, leveraging Philips' iconic logo while creating unmistakable visual contrast. Purple meant AI-generated content requiring physician review. After validation, information turned black and became official record. This let physicians maintain eye contact with patients instead of typing into screens.

Takeaways

The Assistant won the IF Design Award 2024 and is currently being tested in controlled environments with select clients before scaling to critical care contexts. Doctors don't want AI making decisions. They want more time with patients. We're giving them both.

About

Other work

Linkedin

Back

Company

Philips

Year

2024

EMR AI Assistant: Building Trust

in Clinical AI

Context

Doctors were spending over half their workday on administrative tasks rather than patient care. The challenge wasn't just building AI-powered documentation for the Tasy EMR. It was earning clinicians' trust in a tool that would listen to consultations and auto-populate medical records. Usability testing with real physicians revealed immediate resistance: they didn't trust AI in clinical contexts.

My Role

As Product Designer for the Tasy EMR AI Assistant over two years, I was responsible for the voice assistant interface, AI design guidelines, and ambient documentation UX.

I led usability testing with physicians and designed the visual language that would differentiate AI-generated content from validated records.

Solution

One insight changed everything: doctors accepted AI as a copilot when validation stayed in their hands. We tested multiple visual approaches, starting with blue elements that disappeared against Philips' brand colors. The breakthrough came with the purple star, leveraging Philips' iconic logo while creating unmistakable visual contrast. Purple meant AI-generated content requiring physician review. After validation, information turned black and became official record. This let physicians maintain eye contact with patients instead of typing into screens.

Takeaways

The Assistant won the IF Design Award 2024 and is currently being tested in controlled environments with select clients before scaling to critical care contexts. Doctors don't want AI making decisions. They want more time with patients. We're giving them both.