Strategy question

You inherit a signed public-sector AI/data deployment that has not reached production three months after contract start. Security accreditation is incomplete, data interfaces are unstable, and the customer has no agreed definition of 'production-ready.' How would you structure the path from contract to production, what milestones would you use, and how would you surface the top risks early enough to change the plan?

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

Whether you can bring structure and risk visibility to a stalled, high-stakes public-sector deployment without clear success criteria.

Start by defining 'production-ready' jointly with the customer in writing, since the absence of an agreed definition is itself the root blocker.
Audit the current state across three tracks: security accreditation status, data interface stability, and any remaining functional gaps, each with an owner and a real timeline.
Set milestone-based checkpoints (for example accreditation submission, interface stability testing, user acceptance testing) rather than a single go-live date, so progress is visible in stages.
Surface the top three risks explicitly to your own leadership and the customer early, for example accreditation delay being outside your control, so the plan can be adjusted before it becomes a crisis.
Build a weekly risk and status review with the customer that shows real progress against milestones, rebuilding trust through visible, incremental delivery.
Escalate internally for resourcing help on the highest-risk track, likely accreditation, since that is often the longest and least controllable dependency.

Names the root cause precisely: no agreed definition of production-ready, and treats fixing that as the first concrete action, not just a symptom to work around.
Structures the plan into three parallel tracks with named owners rather than one undifferentiated backlog.
Surfaces risk early to leadership rather than hoping it resolves, showing real judgment about escalation timing in a public-sector context.
Shows how milestone-based, visible progress rebuilds a stalled customer relationship.

Jumping straight into technical fixes without first establishing an agreed definition of done with the customer.
Treating security accreditation as a background task instead of a named, tracked risk with an owner.
No plan for surfacing risk early enough to change course before the situation worsens.

How would you handle it if the customer disagrees with your proposed definition of production-ready?
What would you do if accreditation timelines are outside your control and slip repeatedly?
How would you keep your own leadership informed without creating alarm before you have a plan?