Generative artificial intelligence (GenAI) is transforming organizational knowledge work, yet its effects in ill-defined problem domains remain poorly understood. This dissertation examines two such domains: software development and creative ideation. The first essay develops and validates a system dynamics model of how GenAI use affects software developer skills over time. The model draws on theories of metacognitive regulation and cognitive offloading and is grounded in interviews with 30 software developers. Simulations show that GenAI use can produce immediate productivity gains while eroding computational thinking and technical skills of software developers over time. Skill risks are greatest for junior developers and can impede progression toward senior expertise. Policies preserving prompt drafting and refinement and encouraging learning-oriented use of GenAI can mitigate skill erosion. The second essay examines how GenAI assistance in problem formulation and solution generation stages of creative ideation impacts the creativity and diversity of resulting ideas through a 2 x 2 randomized lab experiment. Providing GenAI assistance in problem formulation increased the usefulness of problem statements but reduced their novelty and diversity, whereas GenAI assistance in solution generation increased solution novelty without reducing diversity significantly. Together, the essays show that realizing GenAI’s organizational value depends on designing human–GenAI collaboration to preserve human engagement in problem framing, refinement, and learning, while assessing not only immediate performance but also the longer-term and collective consequences.