Interview Questions › AI Product

An AI companion for lonely elderly users

As a PM at Anthropic, design an AI well-being companion for elderly users that reduces loneliness without creating unhealthy dependence.

Context

Many older people live alone. An always-available AI could help, but could also replace human contact, mislead users about what it is, or miss signs of a medical or mental health emergency.

Clarifying questions to ask first

• Who pays: users, families, care providers?
• Voice, text or a device?
• What happens in an emergency?

How to approach it

1. Start from the user problem, not the model: is AI the right tool here?
2. Describe the experience and where the model fits in it
3. Define what "good output" means and how you would evaluate it before and after launch
4. Name the failure modes (wrong answers, bias, privacy, cost) and how you contain them
5. Close with a launch plan and success metrics

Framework: User and problem -> Is AI the right tool? -> Experience design and human-in-the-loop -> Quality bar and evaluation -> Risks, cost and latency -> Staged launch and metrics

Sample answer

• The biggest risk is over-engagement replacing real relationships
• Success is more human contact, not more time with the AI
• The AI is open about being an AI and nudges towards real-world actions

1. North Star: share of users reporting less loneliness and more human contact
2. Cap engagement metrics instead of maximising time spent
3. Guardrails: escalation accuracy, dependency signals, user wellbeing surveys

A voice companion that is honest about being an AI, nudges users toward family and activities, and escalates to humans when needed. Measure loneliness and human contact, cap time spent.

What interviewers look for

• Outcome over engagement
• Dependency and honesty guardrails
• Escalation to humans
• Family and caregiver ecosystem