Sponsored
Teaching AI to run with the turbines
As energy companies push AI deeper into industrial operations, success increasingly depends on governance, trusted data, and systems designed to augment human expertise, says Andrew Melouney, vice president for digital at Woodside Energy.
In 2026, artificial intelligence is no longer a novelty in the energy sector—it is a necessity. From predictive maintenance on wind turbines to optimizing output from solar farms, AI is being woven into the fabric of critical infrastructure. Yet scaling these systems safely and effectively requires more than just powerful algorithms. According to Melouney, real-world success hinges on three pillars: rigorous governance, high-quality data, and a clear focus on human-machine collaboration.
Woodside Energy, based in Australia, has integrated machine learning models into its operational workflows. These systems help forecast equipment failures, reduce unplanned downtime, and improve production efficiency. AI models analyze sensor data from turbines and pipelines in real time, alerting field operators to potential issues before they cause disruptions.
But with increased automation comes increased responsibility. Melouney emphasizes that governance is foundational. Without well-defined policies for data access, model validation, and explainability, AI can introduce risks that outweigh its benefits. He points to the need for "trusted data pipelines"—curated datasets that are regularly audited and verified for accuracy. In 2026, many companies still struggle with fragmented, siloed data. Overcoming that requires a defense-in-depth approach: standardized data formats, metadata ontologies, and organizational subject matter experts (SMEs) who can act as stewards for information quality.
Equally important is designing AI systems to augment, not replace, human decision-making. Melouney argues that the best models are those that present actionable insights—for example, predicting spare part failures with enough lead time for intervention—rather than autonomous actions. This preserves human oversight in safety-critical environments such as oil rigs and gas plants.
Looking ahead, Woodside is exploring generative AI for operational reporting and knowledge management. While promising, such tools require careful deployment guardrails to ensure they produce factual, context-aware outputs. The company’s experience underscores a broader industry shift: as of mid-2026, nearly 60% of energy firms have deployed at least one AI system in production, per industry surveys, up from 35% in 2023. However, the true competitive differentiator is no longer who adopts AI first, but who manages it best.
In Melouney's view, the goal is not to hand over control to algorithms but to create a symbiotic system where humans and machines learn together. "We're teaching AI to run with the turbines," he says, "not instead of them."