Product design question

How would you improve Google Assistant?

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 product improvement thinking on a mature AI product, requiring a specific, high-value gap rather than a generic wishlist.

Identify a real gap: multi-step, multi-app tasks, like booking a specific restaurant table, still fail or require falling back to a phone screen.
Segment on users who rely on voice hands-free, drivers and users with accessibility needs, where screen fallback isn't a real option.
Identify the pain point: the assistant handles single commands well but breaks down on tasks needing multiple confirmations or third-party app handoffs.
Propose deeper task completion for a few high-value flows, restaurant booking, ride-hailing, and calendar scheduling, with reliable multi-turn confirmation.
Define success as task completion rate for these flows without screen fallback, and repeat usage of voice-only completion.

Names a specific, real limitation, multi-step task failure, instead of a vague make-it-smarter answer.
Prioritizes the accessibility and hands-free use case, where the stakes of screen fallback are highest.
Picks task completion rate as a concrete, testable success metric.

Giving a vague answer like make it smarter or more like ChatGPT without a concrete mechanism.
Ignoring the hands-free and accessibility use case, where this gap matters most.

Which specific task would you prioritize building first?
How would you measure whether users trust the assistant enough to complete tasks hands-free?