What the work actually is
You write sales management problems hard enough that a strong model gets them wrong, then you judge the answers. A typical item might be a distorted pipeline where three reps have stuffed late-stage opportunities before quarter close, a territory redesign that strands an account with the wrong rep, a comp plan whose accelerator invites sandbagging, or a coaching decision about a rep hitting quota entirely through one inherited account. You supply the scenario, the numbers, the defensible answer, and the reasoning behind why the obvious answer fails. Then you review model output and write feedback specific enough that an engineer who has never carried a bag understands where the reasoning broke.
What the screening looks for
The screen is AI-led and pushes on specifics. Expect to be asked what you managed — headcount, quota, sales cycle length, ACV, motion (inside, field, channel) — and then asked follow-ups that only someone who actually ran the forecast call can answer. Retail store management does not count here; the listing is explicit about non-retail B2B, wholesale, or commercial environments. The second thing measured is evaluation judgment: whether you can distinguish a model answer that is wrong from one that is merely different from how you would have handled it, and whether you can state a grading rationale rather than a gut reaction.
Logistics
- Fully remote and asynchronous; no standing meetings, work claimed from a queue
- Contract engagement, typically part-time; contributors commonly report 10–20 hours a week
- Observed rates $75–110/hr, generally set by demonstrated depth and how clean your first batch of items is — not guaranteed
- Output is text: scenarios, rubrics, rationales, and written critiques of model responses
The people who last in this work are the ones who enjoyed building enablement material or running rep certifications. If writing a detailed deal-review teardown sounds like a chore rather than a satisfying afternoon, the volume here will wear thin fast.