Over two years, vendors and a credentialing body partnered to develop and validate an AI-enabled automated test form assembly tool. It supports blueprint, psychometric, and content parameters, including semantic similarity for enemy item detection. In two pilots, forms built by psychometricians and automatically assembled forms were deployed in parallel. A third-party analysis found comparable performance between both form types. Adoption followed across multiple stakeholders. This session explores the value of partnerships between technology vendors, psychometricians, and credentialing organizations in piloting and validating AI-driven assessment tools. We’ll discuss best practices and challenges in ensuring reliability and validity in AI applications.