Image description: A computer circuit board with a central black square, with the letters AI prominent in white. Photo by Igor Omilaev on Unsplash

Ofwat has published its first artificial intelligence adoption plan for the water sector, setting out how AI can be deployed across water and wastewater services. The plan arrives as the regulator prepares to hand over to the new regulatory bodies that will replace it under the government’s water sector reforms.

Already embedded

The plan recognises that AI is no longer a future prospect for the sector but part of day-to-day operations. Companies are already using it for leakage detection, network management, asset maintenance, billing, customer support and regulatory reporting, combining established machine learning with newer generative AI tools. Ofwat frames the question as how AI should be deployed, not whether.

The regulator stressed that AI must not replace professional judgement, and that water companies will remain accountable for decisions supported by AI systems. It also pointed to the rollout of 10 million smart meters between 2025 and 2030 as a data source expected to improve leakage identification and demand forecasting.

Five priorities, and a caveat

The plan covers five areas: understanding sector-wide adoption, developing formal guidance, enabling innovation, building monitoring frameworks, and strengthening Ofwat’s own AI capability. Poor data quality emerged as the most commonly cited barrier among companies. Ofwat intends to consult in autumn 2026 before publishing initial guidance, and envisages a joint cross-regulator AI plan by 2027-28, to be carried forward by the new regulators once established.

Ofwat was candid about its position, describing the plan as its best intentions for the work before the new regulators are operational. The document amounts to a transitional blueprint from a body that will be replaced under current reforms.