What the work involves

You receive an image of a business slide — an emissions summary, a Scope 3 inventory breakdown, a corrective action plan, a permit compliance status update, a CSRD or GRI-aligned reporting page — and rebuild it natively in Microsoft PowerPoint. Reconstruction is the baseline; the real task is judgment. You decide whether a stacked bar should become a waterfall, whether a title reading "Emissions Overview" should become an evidence-supported assertion, whether a caveat about Scope 3 estimation methodology belongs in the footnote or the chart itself. You also decide what to leave alone.

Every slide ships with a detailed log: what you changed, why, what you considered and rejected, and which original choices you deliberately preserved. The log is often more valuable to the model than the slide. A rebuilt slide with a thin rationale is a weak deliverable; a modestly edited slide with a sharp explanation of why the original was already defensible is a strong one.

What the screen looks for

  • Domain fluency under follow-up. Expect probes on Scope 1/2/3 boundaries, market-based versus location-based accounting, materiality, assurance language, and the difference between a regulatory definition and a colloquial one. Vague sustainability vocabulary unravels quickly.
  • Restraint. Candidates who improve everything score worse than candidates who can articulate why a chart, caveat, or hedge was correct as drafted.
  • Factual integrity under design pressure. Simplifying a chart must never quietly drop an uncertainty band, a baseline year, or a boundary qualifier.
  • Real PowerPoint execution. Native shapes and charts, not screenshots dropped into a frame.

Logistics

Fully remote, contractor, async. Work arrives in sprints, so volume is uneven — heavy weeks followed by quiet ones. Each production assignment is scoped to 1–2 hours and is expected to be delivered within that window. Observed pay on this listing runs $150–350/hr, set by the platform and not guaranteed; rates typically track demonstrated depth and calibration performance rather than tenure alone. No prior AI experience is required.