What the work actually involves
You'll be handed prompts that mirror the decks you already build: a three-slide business case for a security platform, an executive summary for a renewal at risk, a QBR that has to open with bad usage numbers. Some tasks ask you to produce the deck yourself as a gold-standard reference. Others give you a model-generated version and ask you to score it, mark exactly where the logic breaks, and rewrite the failing sections. The critique is usually worth more than the artifact — a rating with no reasoning attached is close to useless to the lab.
The evaluation vocabulary borrows heavily from consulting: horizontal logic (do the action titles read as a coherent argument on their own?), vertical logic (does each slide's body actually support its title?), MECE grouping, so-what testing on every exhibit. If you've never worked that way explicitly, you have probably done it implicitly in every deck that survived a VP review. You'll need to make it explicit in writing.
What the screen is looking for
- Real AE tenure at a B2B or SaaS company with a recognizable sales motion — not adjacent marketing or CS roles relabeled
- The ability to articulate why a slide fails, in specific terms a non-salesperson could act on
- Genuine PowerPoint craft: layout discipline, exhibit design, knowing when a chart should be a table
- Written English that holds up under scrutiny, since the written artifact is the deliverable
An AI voice screen sits at the front of the process. It asks about your deals, your decks, and your availability, and it follows up on vague answers. Expect it to push on specifics.
Logistics
Fully remote and fully asynchronous. Five to twenty hours a week is the stated range, and contributors who want more volume can usually get it. There are no live calls, no accounts, no quota. Pay has been observed at $100/hour; rates on this kind of work vary by task type and are set by the lab, not guaranteed by the listing. Most people fit it around a full-time sales job.