We develop a partial-equilibrium theory of credentialed employability in accounting. An environment in which skills are valued not only for task production but also for auditability and standards-contingent defensibility under oversight. We distinguish routine procedural competence (𝑇𝑟), advanced accounting judgment (𝑇𝑎), behavioral capability (𝐵), and defensibility/standards-accountability competence (𝐷). Generative AI reduces the marginal production value of routine procedure but can increase the marginal value of defensible judgment because human sign-off, documentation, inspection risk, and litigation exposure remain human-assigned. Universities choose curricula subject to adjustment costs and a routine floor imposed by licensure/accreditation coverage. Employers choose hiring and complementary investments under liability-weighted productivity. The model provides accounting-specific predictions (i) GenAI shifts returns away from routine procedure except where credentialing binds (ii) GenAI increases the joint return to advanced judgment and behavioral capability; and (iii) competency frameworks decay faster when governance is fragmented across universities, professional bodies, firms, and regulators, especially when oversight and credentialing update on different clocks. We conclude with implementable measurement strategies using institutional data (exam blueprints, inspection intensity proxies, and standards volatility).