Ofwat Bets on Algorithms to Fix Old Pipes

The regulator's new strategy backs wider AI use in leakage detection, overflow prediction and demand forecasting.

18 Jun 2026

Ofwat has decided that the question is no longer whether water companies should use artificial intelligence, but how. Its June strategy for AI adoption during the sector's transition period accepts that algorithms already sit inside leakage detection, network optimisation, billing and regulatory reporting. What the regulator wants now is accountability to match the ambition.

That caveat matters more than it might sound. Ofwat insists AI must reshape roles rather than replace judgment, and that water companies remain answerable for whatever decisions the technology informs. Poor data quality, inconsistent telemetry and patchy granularity, the regulator concedes, are still the biggest obstacles standing between utilities and the value AI promises.

Wastewater offers the clearest test case. Machine learning models already predict sewer blockages, pollution incidents and storm overflow risk, echoing deployments such as Southern Water's smart sewer system in Whitstable. A rollout of ten million smart meters between 2025 and 2030 should sharpen leakage identification and demand forecasting further still.

Timing complicates the picture. The strategy lands as a government white paper proposes folding regulatory functions into a single new body, leaving open exactly who will enforce AI governance once the current structure is dismantled. Utilities investing in these systems now are, in effect, wagering on rules that have yet to be written.