The work

You'll receive AI-generated work products of the kind a sales leader or GTM operator would actually produce: a QBR deck, a competitive battlecard, a pipeline forecast in a spreadsheet, a territory and quota plan, an outbound email sequence, a pricing proposal. Your job is to judge whether the output would survive contact with a real buyer or a real CRO. That means checking claims against evidence, testing whether the funnel math holds, noticing when a deck's narrative structure collapses, and flagging the confident-sounding-but-wrong assertions that models produce most often in commercial contexts.

Evaluations follow a rubric — typically dimensions like factual accuracy, domain rigor, structure, and presentation quality — and each score needs written justification. Slides matter more here than in most evaluation work: you'll be asked to assess layout, hierarchy, chart choice, and whether a deck reads cleanly, so hands-on fluency in Google Slides and PowerPoint is a hard requirement rather than a formality.

What the screening looks for

  • Verifiable commercial experience. Expect probes on the specific motions you've run — SMB versus enterprise, inbound versus outbound, what you carried as a number, what tooling you lived in.
  • Depth under follow-up. The AI interviewer will push a second and third layer on any GTM concept you raise. Vague frameworks-talk reads poorly; specifics about deals, stages, and conversion assumptions read well.
  • Evaluation judgment. Can you separate "I would have written this differently" from "this is wrong"? Reviewers who score on taste alone get filtered out.
  • Written clarity. Feedback is the deliverable. Terse, structured, actionable.

Logistics

Fully remote, asynchronous, hourly. Tasks are drawn from a queue with turnaround windows rather than fixed shifts, so most reviewers fit it around existing work — though throughput tends to be better if you can commit predictable blocks of 8–15 hours a week. Pay bands are as observed on the platform and vary by task type, calibration performance, and project; nothing is guaranteed.