A charity's HR coordinator runs a staff onboarding process that works well overall: the HR manager checks every welcome pack against the charity's policies before it goes out, and new starters rate the process highly. The one slow step is writing a tailored welcome letter for each role, which takes the coordinator about an hour per new starter. How should the coordinator bring Claude into this workflow?
- ARedesign the whole onboarding process around Claude, so that it drafts every document in the pack and sets the order of the onboarding steps.
- BUse Claude to draft the tailored welcome letter from the role details, then pass it into the existing pack for the HR manager's usual policy check. Correct
- CHave Claude draft and send each welcome letter straight to the new starter, so the HR manager's check no longer delays the slowest step.
- DAsk the technical team to connect the HR system to Claude so letters are generated automatically whenever a new starter is recorded.
Why A is wrong: Redesigning can sound more ambitious and therefore more valuable. Here the process already works and only one step is slow, so rebuilding everything adds disruption and risk without addressing a real problem beyond that step.
Why B is correct: The bottleneck is a single drafting step inside a process that otherwise works, so augmenting that step is the proportionate change. The letter still flows through the HR manager's existing policy check, so the control that protects the charity stays in place.
Why C is wrong: Removing a wait looks like a further time saving. It also removes the policy checkpoint the process relies on, so a drafting error or policy mismatch would reach the new starter with no one having reviewed it.
Why D is wrong: Automation can seem like the natural end point for a repeated task. The volume here is one letter per new starter and the coordinator can draft it in the Claude app, so a custom integration is an unnecessary escalation for a need the apps already meet.