Software vendors increasingly seek customers’ operational data to train AI systems, often expanding their rights through quietly revised terms. Customer backlash forced HubSpot to reverse one such change within four days, while Meta’s broader proposal drew congressional scrutiny. Amid this growing data rights conflict, this article offers customer-side counsel a framework for evaluating and negotiating AI data use terms in software agreements, including questions to ask, red flags to identify and contractual safeguards to seek. It draws on insights from practitioners at Corning Inc., Hunton and Squire Patton Boggs, and panelists from Clearly AI, Costco, Integrative Privacy, Paul Weiss and Sunrise Privacy Consulting who spoke at the recent IAPP Privacy.Security.Risk. + AI conference. See “Pain Points and New Demands in AI Contracts” (Jun. 18, 2025).