Utility data sits locked in procurement, asset management, and estimating systems, reachable only via SQL or analyst intervention. With Evergy, this session cuts straight to the technology of building natural language interfaces (NLI) to access data: what works, how it was built and what we learned.
Drawing on deployments with Evergy and other utilities, we walk through the architecture of NLI using utility data, text-to-SQL and hybrid retrieval (RAG across structured and unstructured sources) over a governed semantic layer, with LLM orchestration and query validation tying it together. We will share honest assessments of where each piece earned its place.
Not a vendor pitch or theoretical overview, a practitioner-led, example-driven discussion grounded in real utility deployments, with Evergy sharing the utility perspective.
Session Takeaways:
• Architectural patterns for NLI solutions using structured data
• Lessons on accuracy, governance, and integration
• A framework for piloting NLI in your organization
