Ever since Generative AI (in my case, M365 and related Copilots) started to gain traction within organisations, several ideas around how enterprises can effectively deploy M365 Copilot have been at the forefront of my mind. Although I touched on this in last week's podcast (look for Cloudy with a Chance of Insights on YouTube and/or Spotify if you are interested), I thought it worth expanding on those in a more formal setting. In this article, I attempted to identify a series of pillars organisations should consider before deploying Generative AI workloads. Please don't hesitate to suggest additional/alternate actions or disagree/expound upon those that I have highlighted.