AI at OG&E — S.E.E. DOCC Committee Presentation
Co-presented OG&E's AI methodology at the Southeastern Electric Exchange peer-utility forum.
The talk
In April 2026, the Southeastern Electric Exchange (S.E.E.) DOCC Committee convened distribution-operations leaders from member utilities across the southeastern U.S. for the Spring Meeting. The agenda invited each contributing utility to share what they were doing with emerging operational technology. OG&E’s slot was AI.
I co-presented “AI at OG&E — Transforming Grid Operations — from the media wall to the switch order” — a 13-slide interactive deck that walked the room through three production applications built using the seven-step AI-assisted development methodology.
The presentation embedded live demos, mermaid-diagrammed architectures, and a frank conversation about the development pace, tooling choices, and governance posture that made the work possible.
What the deck covers
- Title — context: S.E.E. DOCC Spring Meeting, April 2026.
- About OG&E — 910K customers, 5,500 transmission miles, 55,500 distribution miles, 542 substations (95% SCADA-controlled), OSI ADMS with FLISR and SOM.
- Outline — the order of what’s coming.
- The Platform: Nighthawk — the in-house web platform hosting every app shown.
- Project 1 — The Media Wall — DCC media-wall redesign with AI.
- Project 2 — Resource Availability — workforce-scheduling app with measured ROI tied to CMI / SAIDI.
- Resource Availability — Process Comparison — before / after operational diagrams.
- Resource Availability — What It Does — features and value.
- Project 3 — Operator Switching Dashboard — making switching quality measurable.
- What AI Did On These Projects — the role AI played in design, development, and integration.
- How We Use AI — End to End — the seven-step methodology (align → prompt-engineer → analyze → prototype → review → approve → deploy).
- The Journey Since October 2025 — timeline of AI-first development inside a utility.
- Questions? — contact slide.
The seven-step methodology
Align with stakeholders → Prompt-engineer with AI → Analyze data with AI → Build prototype with AI → Stakeholder review → (Loop if not approved) → Deploy to Nighthawk
The cycle has been used to ship three production AI applications at OG&E. The compounding part: every cycle produces two outputs — a better application and a sharper sense of how to use AI well on the next one. The process itself improves with each project.
Why it landed
Three concrete reasons the deck resonated with peer utilities:
- It showed shipped work, not slideware. Each project demonstrated had production code, deployed users, and measurable outcomes.
- It named the methodology. Most utilities are still figuring out how to do AI. A repeatable cycle gives them a framework to adopt or adapt.
- It was peer-credible. The conversation was utility-to-utility, not vendor-to-customer. Same operational constraints, same regulatory environment, same workforce realities.
Take a look
The full deck is embedded below. Use arrow keys or the on-screen controls to navigate.