Product design question

Harvey wants to launch a new AI product for legal teams, but customer needs are still ambiguous. How would you define an MVP that solves one concrete workflow end-to-end, respects enterprise data and security constraints, and creates value for both practitioners and economic buyers? Be specific about the target user, job to be done, scope cuts, and what you would defer.

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 MVP scoping for an ambiguous legal AI product, balancing enterprise data constraints against creating value for both users and buyers.

Pick a specific target user, for example a mid level associate doing contract review, rather than legal teams broadly.
Define one concrete job to be done, such as flagging non standard clauses against a firm's playbook, with a clear before and after.
Scope cuts explicitly: defer multi document comparison and custom playbook authoring to a later release, ship against a single standard playbook first.
Bake in enterprise constraints from day one, such as data never leaving the tenant and full audit logging, since these are launch blockers not add ons.
Design the value story for both the associate, time saved, and the economic buyer, risk reduction and billable hour protection.
Validate the MVP with a design partner before general availability to confirm the scope actually solves the job to be done.

Names a specific user and workflow instead of a generic legal AI pitch, showing the candidate can narrow ambiguity into something testable.
Treats security and data constraints as part of the MVP definition, not a phase two concern, given the enterprise legal context.
Explains value for two different stakeholders, the practitioner's time savings and the buyer's risk reduction, since legal sales require both.

Defining an MVP so broad it tries to serve every legal workflow at once.
Treating security and compliance as post launch hardening instead of a core scope requirement for legal enterprise buyers.

How would you validate the MVP is solving a real job before committing engineering resources?
What would you cut first if the timeline got compressed by half?