The work
This is comparative evaluation rather than legal advice. You receive a prompt set drawn from the kinds of questions an in-house lawyer actually fields — contract interpretation under the Code civil, commercial terms and conditions générales, clause drafting, distribution and supply arrangements, liability and termination questions — and run each prompt through two AI platforms. You then read both answers closely, decide which is stronger and why, and score each against a rubric supplied by the client. Every evaluation ends with concise written feedback: what the model got right, where it drifted from French law, whether it hallucinated an article number or imported a common-law assumption, and how a partner-level reviewer would rate its usefulness to a real GC.
The judgment being tested is a lawyer's, not a proofreader's. A fluent, well-structured answer that cites the wrong article or misstates the régime of a clause should score below a plainer answer that is correct. Consistency across items matters as much as accuracy on any single one — reviewers who apply the rubric the same way on item 40 as on item 1 are the ones who get invited back.
What the screen looks for
- Verifiable practice history: where you were admitted or trained, which companies or sectors you supported in-house, and the categories of matters you actually owned.
- Depth under follow-up: the AI interviewer will push past a general answer on French commercial and contract law to see whether the specifics hold — expect follow-ups on provisions you cite.
- Rubric discipline: whether you can separate presentation from substance and defend a score rather than a preference.
- Written precision in the language you'll be reviewing in. Feedback must be terse and unambiguous.
Logistics
Fully remote and asynchronous — work is claimed and submitted through Mercor's platform on your own schedule, with per-batch deadlines. The initial engagement is scoped at roughly 10 hours, and reviewers who deliver clean, consistent work are commonly extended onto further batches. All work is under NDA. The observed rate for this listing is USD 150–160 per hour; pay bands on Mercor reflect what has been posted for the project and are not a guarantee of ongoing volume.