You inherit an architecture review board with a six-week queue and a poor reputation with engineering. What do you change first?

What the interviewer is testing

Whether you can rebuild a governance function's legitimacy rather than optimising its throughput.

The diagnosis

A six-week queue means the board is on the critical path for far more than it should be, and a poor reputation means teams experience it as an obstacle rather than as help. Both follow from the same cause: it is reviewing things a machine should check, and arriving too late to change anything.

What I would change, in order

1. Automate the rule-shaped checks. Region, encryption, tagging, approved technology, network exposure, mandatory telemetry — as policy code and guardrails. In most boards this is half the submissions, and removing it is immediate relief with no loss of control.

2. Tier by consequence and reversibility. Reversible and local: no review. Hard to reverse or crossing boundaries: lightweight asynchronous review — an ADR circulated with a comment period, no meeting. Organisation-level commitments: the board.

3. Move earlier and reframe as help. Engage at design time, and make the reviewers' job to improve the design and surface risk rather than to approve. Where sign-off is genuinely required, separate it from the improvement conversation, because the two produce different behaviour in the room.

4. Give teams a self-service pre-check so they can validate against the automated rules before submitting anything, removing the round trip for compliant designs.

5. Publish decisions and reasoning, so teams can learn the standards rather than discovering them in review.

The reputational repair

The fastest way to change how engineering perceives the function is for the board to be visibly useful: help a team through a hard design problem, and let that team say so. Governance functions earn legitimacy by contribution, not by authority.

What a strong answer adds

Measuring the board: time in queue, proportion of submissions that resulted in a material change, and findings raised at design stage versus at launch. A board that cannot show it changes outcomes is performing ceremony, and measuring it is what makes reform possible.

Common weak answers

Adding reviewers to clear the queue. Meeting more frequently, which scales the ceremony.