Generative AI is already transforming creative and language-based tasks. Freely available LLMs and prompt engineering are often all you need. However, for precise, repeatable tasks like data analytics, it’s a little more complex.

Casey’s Analytics Center of Excellence was looking for ways to free up analysts for strategic efforts, reduce last-minute requests, and speed up time to insights for our partners. Could a GenAI assistant, tailored to our business, conduct descriptive and diagnostic analyses for business users at their direction? We decided to find out.

In this presentation, we’ll discuss why it was the right time to explore GenAI for data analytics, including the right use case, leadership buy-in, robust data, capable tech stack, and strong vendor partners. We’ll share how we secured stakeholder support, the value we aimed to realize, our proof of value approach, and our use of metric-driven development (MDD).

We’ll also cover the general and specific challenges of GenAI in analytics, recent technology advancements addressing these hurdles, and how we envision GenAI agents fitting into our organization. Finally, we’ll present real-world interactions and outputs (randomized/edited to protect Casey’s data), share lessons learned, and outline our future plans.