Estimation question

Estimate the cost savings of routing 10% of an engineering org's tasks to Devin.

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

Estimation combining organizational sizing with AI-adoption assumptions to reach a defensible cost-savings figure.

State the goal: estimate annual cost savings from routing 10 percent of tasks to Devin for a mid-size engineering org.
Assume an org size, for example 1,000 engineers, each costing roughly 200,000 dollars loaded annually.
Assume each engineer spends a portion of time, say 20 percent, on tasks suitable for delegation like bug fixes, tests, and migrations.
Compute delegable task-hours: 1,000 engineers times 20 percent times full-time equivalent equals 200 FTE-equivalents of delegable work.
Apply the 10 percent routing rate: 20 FTE-equivalents worth of tasks move to Devin, valued at 20 times 200,000 dollars equals 4 million dollars in engineer time.
Net out Devin's cost (a subscription or usage fee) and a productivity discount, since AI-completed tasks may take longer to review, to reach a net savings estimate, clearly flagged as assumption-driven.

Builds the estimate top-down from org size and delegable task share rather than guessing a number outright.
Nets out the AI tool's own cost and review overhead, not just gross engineer-time saved.
States every input as a labeled assumption so the interviewer can challenge any one of them.
Sanity-checks the final figure against Devin's actual subscription pricing to see if the value proposition holds up.
Notes that the 10 percent figure itself should be validated by a pilot before being used for a business case.

Skipping the review-overhead discount, overstating net savings.
Using an arbitrary org size or cost-per-engineer without stating it as an assumption.

Chapters of the AI PM course that teach what this question tests.