Design AI features for Google Docs (writing, reviewing, and summarization).

### Signal to interviewer

I can design AI that supports collaboration quality, not just single-user drafting speed.

### Clarify

I would clarify document types, collaboration patterns, and whether users optimize for speed, quality, or governance.

### Approach

Use a lifecycle loop: guided writing, rationale-based review, and intent-specific summarization with source anchoring.

### Metrics & instrumentation

Primary metric: time to collaboration-ready draft. Secondary metrics: suggestion acceptance, comment resolution speed, and summary usefulness ratings. Guardrails: incorrect factual rewrites, voice drift complaints, and review churn.

### Tradeoffs

Stronger auto-rewrite can improve speed but reduce authorship confidence. Strict safeguards improve trust but may lower perceived intelligence.

### Risks & mitigations

Risk: summary distortion; mitigate with citation links. Risk: reviewer overload; mitigate with priority-ranked suggestions. Risk: style mismatch; mitigate with customizable tone profiles.

### Example

In policy drafting, AI flags ambiguous clauses, proposes alternatives with legal tone, and generates executive and implementation summaries from one source doc.

### 90-second version

Design Docs AI as a lifecycle partner for writing, review, and summarization with grounded suggestions, configurable control, and collaboration-aware metrics.