This panel had over 85 attendees and nearly 2 hours of content from our conversation-style discussion and post-event networking groups.
Four practitioners cut through hype in a Grid Innovation Hub panel on using gen AI in plants and across the grid.
The throughline was simple.
Use AI as a force multiplier, not a magic brain.
Start where time and talent are scarce, keep operators in the loop, and build security in from day one.
Watch the complete recording or read the summary below to learn how these operators and founders are using AI to enhance power generation across operations, interconnection, and cybersecurity.
First, meet our panelists:
Amit Patel, Managing Director & Utility Solutions @ Integ
Michael Ryan, Strategic Energy Consultant, Smart Grid & DOE Grant Initiatives @ Kit Carson Electric Cooperative Inc.
Tom Nudell, CEO @ PIQ Energy
Brandon Miller, Regulatory Compliance Specialist - NERC Consultant @ TRC Companies Inc.
Before you get into the summary…
Share GridInnovationHub.com with your peers, colleagues, and friends.
Use AI To Unstick Bottlenecks
Start inside planning, analysis, and reporting loops now.
The planning and interconnection stack is a sweet spot. Let AI orchestrate deterministic tools, not run the grid.
Think agents that kick off standard studies, assemble evidence and draft reports while humans review the steps and the outputs.
This keeps results auditable, speeds cycle time and avoids “black box” control.
Automate power-flow and stability study runs by having AI call known endpoints, not invent code.
Generate first‑pass model validation notes and study reports, then route to an engineer to sign off.
Use coding assistants to scaffold scripts and test harnesses so a small team ships faster.
Security And Compliance Are The Starting Blocks
Treat regulatory standards as the floor, design for resilience.
Security‑first beats penalty‑avoidance.
Treat compliance as a baseline and design the system you’d want before the next audit or incident. Segment anything that touches OT, keep humans in the loop for safety‑critical actions and assume regulation will lag practice for a while.
Ask three things up front: what systems does this touch, who trained it and on what data, and what happens when it fails.
Air‑gap or strictly segment AI that can reach control networks, and scrutinize every integration or API.
Prove value in sandboxes with clear rollback, then widen scope once operators trust it.
Make It Work At The Team Level
Bottom-up beats big generic enterprise AI.
Executives love data lakes. Operators need time back.
Real gains come when analysts and supervisors use AI to build the small tools they would have begged IT to build. Let them express steps in plain language, turn that into software with tests, and review the output like any other work product.
Expect a mindset shift and teach the “dialect” of prompts.
Capture compliance and outage artifacts as events happen so reports aren’t rebuilt weeks later.
Remove duplicate entries across OMS, GADS and TO deliverables with lightweight automations.
Let AI draft 80–90% of routine memos and checklists, then have humans edit for accuracy.
Measure gains in cycle time and error rate, not vanity dashboards.
Keep Humans In The Loop, Deliberately And Often
Playbooks, forecasting, and community-facing communication
Resilience work benefits from AI that assembles signals and suggests actions while people decide.
Blend weather and fuels forecasts, ignition detection, imagery and sensor data to plan PSPS zones, microgrid islanding and battery dispatch.
Standardize messy outage notes so machines can learn from history. Then write playbooks that help crews, dispatch and comms move in sync.
Turn safety manuals into local runbooks that are searchable and stepwise.
Pre‑position crews using forecast‑driven risk maps rather than gut feel.
Require human confirmation for reclosing, islanding and other high‑risk actions.
Clean and structure historic outage records so prediction engines stop tripping on free‑text.
What To Pilot In The Next Quarter
Small scope, measurable ROI, operator trust built-in
Pick one feeder, one plant unit or one study type.
Bake in security reviews on day one. Aim for a win that staff will defend in the next budget meeting.
Study autopipeline: AI triggers a standard interconnection study, logs each step and drafts the report for engineer review.
Active compliance capture: auto‑collect artifacts during outages and events to cut weeks of after‑the‑fact paperwork.
Cyber anomaly canary: a read‑only detector on OT mirrors that flags weird traffic and proposes a playbook step.
Operator copilot: a chat layer that answers “where, what, why” from trusted data, never from live control.
Bottom line: use gen AI to free experts, not replace them.
Make the work auditable, the wins obvious and the risks bounded.
That’s how you get safer, more reliable and more efficient power without the drama.
Thanks for reading, here are some more grid technology resources for you to check out:
👉 Watch the whole 60-minute panel event here.
👉 Find more grid technology panels, meet-ups, and events here.
👉 Register for our networking coffee chats here.
And if you like our content, check out some of these newsletters we love:
