Enterprise Tech with Fexingo · 2026-07-17 · 9 min
Episode 116 of Enterprise Tech digs into one of the hottest contract clauses in 2026: AI model change controls. Lucas and Luna walk through a real negotiation between a Fortune 500 bank and a major AI vendor, where the bank demanded 90 days' notice before any model update - and a rollback option if performance degraded. They break down why standard AI-as-a-service contracts leave buyers exposed to silent drift, how the negotiation shifted from a one-page addendum to a five-page schedule, and what the final terms looked like. Along the way, they touch on the bank's specific concern: a credit-underwriting model that could silently change how it evaluates small business loans, triggering regulatory risk. The episode also covers the vendor's pushback - arguing that model updates are essential for security and accuracy - and how the two sides found middle ground with a testing period and a kill switch. If you're a procurement lead, CTO, or legal counsel negotiating enterprise AI deals, this episode gives you the exact language and leverage points to bring to the table.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
Why Fortune 500s Are Rethinking AI Model EOL Policies62 / 100How Fortune 500s Negotiate AI Agent Software LicensesWhy Fortune 500s Are Rewriting AI Agent Usage PoliciesFortune 500s Rewrite Software Licenses for AI AgentsHow Fortune 500s Negotiate AI Model Output Escrow
