AI could help renewable-energy teams preserve expert knowledge as experienced workers retire, argues Delfos Energy co-founder Guilherme Studart. His example is a wind farm where engineers investigated 28 anomalies and completed corrective work in about 45 days, with projected avoided losses of 9,629 MWh over a year.
Watch Desk analysis
What happened
In a Forbes Council essay, Studart describes four levels of trust in AI for energy operations: spotting an anomaly, diagnosing its likely cause, recommending an action for human approval, and acting within boundaries set by people. He argues that detection alone does little to ease the pressure on scarce experts if they still have to do all the detective work.
At the wind farm in his account, the flagged issues included rising temperatures in generators, gearboxes and transformers, as well as miscalibrated nacelle anemometers. The system did not dispatch work automatically: engineers investigated the signals and worked with the maintenance team to decide what to fix and when. Studart says the corrections were completed in about 45 days and were expected to avoid 9,629 MWh in losses over a year.
Why it matters
The practical distinction is between an AI that adds another alert to an already busy dashboard and one that helps experts move from signal to diagnosis. In energy operations, that could make scarce engineering time go further without handing consequential decisions to software.
The example is Studart’s account of work at a farm served by his company, not independent proof of a general result. Still, it gives readers a concrete way to test the pitch: what did the system identify, what did people decide, and what outcome was projected?
Our read
“AI will replace the experts” is the easy story. Studart makes a more useful case for capturing what experienced technicians know and using AI to extend their reach. The four-step ladder is a sensible way to ask how much authority a system really has. Start with the job it helps people do, not the autonomy promised on the box.
What to watch
- Whether the reported energy savings are later measured against actual performance.
- How the four levels are applied across other renewable-energy sites.
- Whether operators keep consequential maintenance decisions under human control.
Discussion spark: Should energy operators keep AI at the level of diagnosis and recommendations, or allow it to take bounded maintenance actions once its results are proven?
Sources and evidence
- AI’s Role In Energy Isn't Replacing Experts-It’s Extending Their Reach – Forbes (5 November 2026, 15:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.