Case study
Research behind AI voice features: privacy, disclosure and adoption
Speech-to-text and AI features raise privacy and trust questions users notice immediately. I led research that shaped disclosure, adoption and roadmap decisions with Product, Privacy, Legal and Compliance.
The situation
Voice messaging in VR depended on speech-to-text and AI features, but the team had inconsistent design patterns, limited foundational research, and open questions about how first-time users perceive privacy and disclaimer notices. Adoption and retention of voice users were the commercial question; regulatory intent was the constraint.
My role
I led the biannual research strategy across voice, speech-to-text, AI, AR/VR and smart communication experiences, partnering with product managers, designers, engineers, data scientists, and the Privacy, Legal and Compliance teams. I also coordinated governance for studies involving AI, voice and sensitive user data.
What I did
- Tested disclosure with real users. Redesigned speech-to-text disclosures so they preserved regulatory intent while increasing user trust and first-time voice adoption.
- Translated research into decisions. Connected findings to quarterly KPIs, roadmap priorities and longer-term strategy, and gave stakeholders direction on privacy and legal concerns.
- Paired behavior with qualitative data to establish behavior-based benchmarks.
- Made the work reusable. Contributed studies to research libraries; the findings became a reference for voice-based design in AR and wearables.
Outcome
Research informed 12 of 21 high-priority initiatives for the second half of 2022. A single UI finding the team had not previously identified increased new voice users by 30%, and the disclosure redesign showed that privacy transparency and adoption can move together.
Where design mattered
This is the project where design research was the governance mechanism: the disclosure was the control, and testing it with users is how we knew it worked.