• Map CTD AI opportunities: Identify use cases across CTD workflows (knowledge retrieval, document acceleration, exception management support, operational insights, coordination/handoff support, client-response acceleration). • Build and ship: Deliver POCs/MVPs using AWS/Databricks/Dataiku/OpenAI, integrating with operational systems, data, and document repositories. • Roadmap and operating cadence: Own a roadmap and drive an execution cadence (demos, stakeholder reviews, adoption checkpoints, impact tracking). • Balance quick wins vs foundations: Deliver immediate workflow improvements while shaping requirements and reusable patterns that will scale as IT infrastructure matures (integration patterns, evaluation/monitoring, orchestration approaches, governance-aligned design). • Vision and transition plan: Define the CTD AI north star and a staged path from POCs to reusable capabilities to scaled delivery. • Adoption-first delivery: Convert solutions into practical assets - templates, prompt kits, role-based guides, and training - so adoption is built in. • Risk/controls awareness: Incorporate appropriate controls for privacy/security and traceability expectations where needed. • Stay current and apply responsibly: Keep up with the rapidly advancing AI landscape and selectively introduce relevant techniques/tooling that are safe, measurable, and adoptable in CTD operations.