What the work involves

You will read AI outputs the way you would read a consultant's deck or a director's turnaround memo — and then say precisely where it breaks. Typical tasks include grading an AI-generated denial-trend analysis against the underlying data, judging whether a proposed cost-to-collect reduction plan is achievable in a multi-entity health system, rewriting a weak strategic recommendation into one you would actually present to a board, and annotating where a model confuses gross vs. net collection rate, misreads a payer contract term, or proposes automation that ignores front-end registration realities. Much of the work is structured: rubric-scored comparisons between two model responses, written rationales, and occasional authored reference answers that become training targets.

What the platform screens for

Mercor's screen is AI-led and leans on verifiable specifics. Expect to be asked what your net collection rate and days in A/R actually were, how many FTEs and how much annual net patient revenue you owned, which EHR and bolt-on RCM platforms you ran, and what a specific transformation initiative delivered — in dollars, not adjectives. Follow-ups go a level deeper than the first answer, so approximations that sound rehearsed tend to surface quickly. Reviewers also look for evaluation judgment: the ability to distinguish an answer that is wrong from one that is merely incomplete, and to explain the difference in writing that a non-clinical annotator could follow.

Logistics

  • Fully remote, asynchronous, no fixed shifts; most contributors work in blocks of a few hours.
  • Typical commitments run 5–20 hours per week, project-dependent and not guaranteed to be continuous.
  • Pay is observed at $162/hr for this listing; rates vary by project, batch, and calibration performance.
  • Engagements are contractor-based and scoped in waves — some experts are onboarded to a single evaluation set, others recur across quarters.
  • Reliable written English matters as much as domain depth, since your rationales are the deliverable.