What the work involves
You will review AI outputs against the judgment you already apply on the job: does a charity care determination follow the hospital's FAP and 501(r) obligations, is a presumptive eligibility call defensible, would this payment plan letter survive a patient complaint or a compliance audit? Tasks typically arrive as discrete items — a generated financial counselling script, a Medicaid screening recommendation, a self-pay resolution plan, a patient balance explanation — and you score them, flag errors, and write a short rationale explaining what a competent counsellor would have done instead. Some assignments ask you to construct realistic scenarios (uninsured trauma admit, underinsured patient at point of service, sliding-scale FQHC case) and grade multiple model responses against them.
The written rationale matters as much as the score. Reviewers who write "incorrect" get filtered out; reviewers who write "this applies a 200% FPL threshold when the stated policy is 300%, and omits the required plain-language notice" become the ones the lab keeps routing work to.
What the platform screens for
- Verifiable operational history: which health system, what volume, what your charity care conversion and POS collection numbers looked like, and what you personally owned versus supervised.
- Depth under follow-up on 501(r), EMTALA interactions, presumptive eligibility logic, and where financial assistance policy and collections practice actually conflict.
- Evaluation judgment — whether you can separate an output that is wrong from one that is merely unlike how you would phrase it, and whether you can rank two flawed responses.
- Writing that is specific and legible to non-clinicians, since your feedback becomes training signal.
Logistics
Fully remote and asynchronous, with work claimed from a queue rather than scheduled. Most contributors commit 10–20 hours per week; some projects run in short bursts tied to a lab's evaluation cycle and then pause. The $135/hr figure is the observed rate on this listing — Mercor sets rates per project and per reviewer, and it is not guaranteed. Expect a paid or unpaid calibration task before volume work is released, and expect assignment to depend on ongoing agreement with senior reviewers.