An analysis of the VALs genAI benchmarking study by Noah Waisberg, CEO, Zuva.Artificial intelligence tools for contract review have made real progress in recent years. Yet, recent findings from the Vals Legal AI Report vendor evaluation re-raise a critical question: how reliably accurate is Generative AI at contract data extraction?In the Vals study, three of the four Generative AI tools reviewed (Harvey, Vincent AI (from vLex), and Oliver (from Vecflow)) missed a pretty standard Most Favored Nation (MFN) clause. (The human lawyer and Thomson Reuter’s CoCounsel both found it.) As someone who has been working at AI contract review since 2011, I found this result surprising – and troubling. Here’s why this matters and what it teaches us about relying on AI for contract review.Why MFN Clauses Are CriticalMFN clauses are contractual provisions that ensure one party receives terms no less favorable than any others. If a company offers a better deal to someone else, it must extend those same benefits to the MFN-holder. While pretty uncommon in most agreement types, MFNs can carry significant risks—particularly in MA situations

Zuva releases Multi-Level Document Classifier via its API Offering