Scope an AI copilot for 6,500 field technicians

    Hard40 minBreak it down · AI rubric, per criterionCFDE
    • clarify before building
    • success metric
    • stakeholder alignment

    The brief

    Vertana Lifts (fictional) installs and maintains lifts and escalators across India, the UAE and Saudi Arabia. It has 6,500 field technicians and about 140,000 units under maintenance contract.

    The COO has called you in after a board review:

    • The callback rate (a second visit to the same unit within 30 days for the same fault) is 14%. The board wants it under 8%.
    • Mean time to repair for a breakdown is 3.1 hours. Contracts with large commercial clients have penalty clauses above 4 hours.
    • The fleet spans 40+ controller models from Vertana and acquired brands. Manuals are PDFs, many scanned; some are only in Italian or Mandarin. Senior technicians keep their own notes in WhatsApp groups.
    • Many machine rooms and pits are in basements with no mobile signal.
    • Lock-out/tag-out and safety procedures are mandatory and audited. A fatal incident in 2024 is still under investigation in one country.
    • Technicians in the UAE and Saudi are mostly from South Asia; preferred languages include Hindi, Malayalam, Tamil and Urdu. The technicians' union in India has raised concerns about "monitoring apps".
    • The COO has a budget but wants something on technicians' phones within 4 months.

    The COO says: "I want a ChatGPT for my technicians. They type the fault, it tells them the fix."

    Break this down. The panel will probe how you handle the safety boundary, the stakeholders and the metric, not just the technology.

    How it's graded

    • Asks the questions that change the solution2 pts
    • Precise metric, baseline and guardrails2 pts
    • Draws a clear safety boundary2 pts
    • Maps stakeholders and adoption1 pts
    • Phased approach that fits the constraints2 pts
    • Realistic risk list1 pts

    Your answer

    What would you ask first, and why? Rank them. Do not propose a solution yet.

    0/220 words

    State your working assumptions, then define the metric precisely: baseline, target, how it is measured, and guardrail metrics.

    0/180 words

    Only now: the options you would consider, tied to what the answers might reveal, and what you would build first.

    0/250 words

    What could go wrong, who could be harmed, and how you would contain it.

    0/150 words

    Graded by AI against the ADaSci rubric. Your answers stay private; only your score shows on your profile.