What the work involves
You receive AI-generated outputs — PA determinations, medical-necessity write-ups, appeal letters, formulary lookups, coverage-determination drafts — and judge whether they would actually survive submission to the named payer. That means checking the drug against the correct formulary tier and step-therapy sequence, confirming the clinical documentation cited supports the criteria being claimed, catching hallucinated policy language or invented J-codes, and flagging where the model conflates medical benefit with pharmacy benefit. Much of the task volume involves specialty categories where criteria are dense: oncology, rheumatology, immunology biologics, infusion therapies, and high-cost orphan drugs.
Beyond scoring, you write the correction. Reviewers are expected to explain why an output fails — which criterion was missed, what the payer actually requires, what the appeal or exception pathway should have been — in structured feedback that becomes training signal. Vague ratings without a defensible rationale are the most common reason contributors get rotated off a project.
What the platform screens for
- Verifiable operational history. Mercor's screen probes specifics: which PBMs and portals you worked in, typical daily PA volume, your approval and overturn rates, the therapeutic classes you handled most.
- Depth under follow-up. Expect the interviewer to push past a general answer into a concrete one — the exact step-therapy sequence for a TNF inhibitor under a commercial plan, or the timeline and evidentiary standard for a Part D expedited coverage determination.
- Evaluation judgment. Can you separate "clinically reasonable but non-compliant with this payer's policy" from "factually wrong"? Both matter; they are scored differently.
- Written clarity. Feedback quality is graded. Rambling or hedged rationales fail calibration.
Logistics
Fully remote and predominantly async. Contributors typically commit 10–20 hours weekly, self-scheduled, with occasional live calibration sessions in US time zones. Work is contract-based and project-scoped: volume fluctuates with the lab's task pipeline, and continuation depends on maintaining agreement scores against gold-standard reviews. The $75/hr figure reflects observed rates for this listing tier and is not guaranteed — bands shift by assessed depth and project.