Beyond the Wrapper: Making AI Work for Data Teams

In this episode of The Dashboard Effect, Brick and Landon give an update on Celeste, the internal AI tool Blue Margin has been building to help data teams query, explore, and visualize client data. They dive into what’s worked, what hasn’t, and the fork in the road between making Celeste valuable for engineers versus business users. Along the way, they share lessons about AI adoption, the limits of “wrappers,” and where emerging tools like Azure’s AI Foundry could take things next. If you’ve ever wondered what it’s really like to build and test an AI tool inside a data company, this conversation pulls back the curtain. Timestamps 0:00 – Welcome and episode intro 0:19 – What is Celeste? Building an internal AI tool 1:51 – Engineers vs. business users: different needs 3:26 – The fork in the road: visuals vs. queries 5:00 – How GPT-5 fits in (and doesn’t) 5:58 – Exploring Azure AI Foundry and agent workflows 6:36 – What’s next for Celeste and final thoughts About Blue Margin Blue Margin helps middle-market companies use their data to create focus, accountability, and growth. We specialize in building modern data platforms and Power BI dashboards that turn information into a competitive advantage. Learn more at https://bluemargin.com