Yorkshire Water launches AI water-quality pilot at 20 inland bathing sites

TL;DR:

• Yorkshire Water has partnered with AI firm UnifAI Technology on a two-year research project to use machine learning for real-time water-quality prediction at 20 inland bathing sites across Yorkshire, including Ilkley, Wetherby and Knaresborough.
• The pilot will gather over 7,800 samples from the 20 sites, combining specialist monitoring sensors and on-site sampling with UnifAI’s prediction models — aimed at closing the lag between sampling and public information that currently leaves swimmers learning about health risks “after the fact”.
• Project partners include the Rivers Trust, the British Standards Institution and Southern Water, signalling cross-utility interest in standardising AI-assisted bathing-water monitoring beyond Yorkshire.

The deployment is one of the more concrete UK utility AI use-cases announced this year and directly addresses a problem flagged in successive Environment Agency reports — that traditional bathing-water testing relies on periodic laboratory work and inevitably reports historic conditions rather than present risk.

Context and Background

The structural gap UnifAI’s chief commercial officer Dan Byles describes — periodic sampling producing a historic picture rather than real-time risk — has been a longstanding criticism of England’s bathing-water regime. Current Environment Agency classifications are produced from samples taken across the bathing season and published with material delay, which makes them useful for site classification but limited for day-to-day swimmer decisions. AI prediction trained on sensor and sample data could in principle close that gap if the models prove accurate enough for public health communication.

The inland-bathing focus is significant. England designated 27 new inland bathing-water sites in 2024 and 2025, including Wharfe at Ilkley, in response to growing wild-swimming participation and persistent campaigner pressure over sewage discharges. The Yorkshire Water sites overlap with that newly designated stretch, which means the AI pilot will also produce useful data for the Environment Agency’s evolving inland-bathing regulatory regime — one of the more politically charged areas of UK water policy following Ofwat’s 2024 enforcement settlements.

Yorkshire Water’s coastal delivery and engagement manager Faye Cossins framed the goal as giving “communities near real-time insights so they can make confident, informed decisions about taking a dip” — pitching the project as community-information service rather than just regulatory-compliance tool. The British Standards Institution partnership is the structurally interesting element: if the methodology can be standardised, the model becomes exportable to other UK water companies facing similar inland-bathing pressure.

Looking Forward

The two-year window means meaningful results will not surface until 2028, but the Rivers Trust and Southern Water involvement suggests other UK water companies are watching for a model they can replicate. Expect Ofwat and the Environment Agency to engage with the BSI-led standardisation work — if validated, the approach could shape how inland-bathing risk communication evolves through the 2026–2031 Asset Management Plan period.