Product design question

What would you build to replace service dogs?

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Tests open ended product design in a physical, safety critical context: reasoning about a hard to automate human and animal task.

Define the specific service dog function to replace, such as guiding visually impaired individuals or detecting medical events like seizures.
Identify why this is hard: service dogs combine mobility assistance, real time judgment, and emotional companionship that technology struggles to replicate fully.
Propose a scoped solution, such as a combination of a navigation robot or smart cane with sensor based medical event detection, rather than a single general purpose replacement.
Address the trust and safety bar explicitly, since failure in this category has real physical safety consequences.
Define success: task completion rate for the specific function, and user trust or reliance over time.
State clearly which specific service dog function you chose to focus on, since the full role is too broad for one product.

Scopes the problem to one concrete function, for example guiding a visually impaired user safely through obstacles, rather than trying to replace all service dog capabilities at once.
Proposes a specific mechanism, a sensor equipped smart cane or small navigation robot combined with haptic feedback, grounded in existing assistive technology approaches.
Addresses the safety critical bar directly, proposing extensive real world testing and a human fallback option before full reliance is expected.
Sets a measurable goal, for example a stated task completion and safety incident rate compared to trained service dogs in equivalent scenarios.

Trying to replace every function of a service dog at once instead of scoping to one specific, addressable task.
Underestimating the safety bar required, since failures in this category can cause real physical harm to a vulnerable user.

How would you validate this is actually safer or as reliable as a trained service dog?
What would you do for the emotional companionship aspect that is much harder to replicate?