The GenAI landscape has rapidly evolved over the last year in how information can be aggregated and acted upon. Particularly dramatic is the shift to “Agentic” workflows that allow work to be done on behalf of the user or enable LLM applications themselves to do increasingly sophisticated multi-stage actions such as research and code generation and review. This talk will work through our journey and perspectives learned around a shift away from isolated chat based tools (eg chatbots) to more integrated tooling and how we are working to optimize value and efficiency for end users while still maintaining governance and oversight over usage.