Discussion about this post

User's avatar
Stuart Patterson's avatar

Great question, fortunately, not as difficult as one might think given user (vs. host) control.

Proactive, voice dialogs can be configured by caregivers at PVAI setup as they are generally (1) reminders, (2) check-ins or (3) offers to engage with content: eg, Don't forget to take your new meds! How are you feeling today? or Would you like to listen to music or play scrabble?. Caregivers can also configure how the PVAI will respond and, if/when needed, "act" based on responses of the care recipient (eg, notify the CG via text if the CR says "I feel horrible!").

As providers of PVAIs and numbers of CG and CR users scale, they'll have multiple options -- both automated options managed by the PVAI based on historical data and options managed by CGs (or their human proxies) asynchronously or in real-time.

Reactive, voice-AI dialogs with CRs will operate as they do now with all voice-AI agents, although our vision is that CGs will be able to set guardrails as suggested in today's NYT op-ed about the movie Her!

Rainbow Roxy's avatar

Couldn't agree more. Your 'humans-at-the-helm' for 24/7 proactive voice AI is so insightful. How do you envision that constant oversight scaling, especialy? A real systems design puzzle.

No posts

Ready for more?