How AI and digital innovation are transforming the UK water industry

It's fair to say that the UK’s water industry is seldom out of the news. From leaking water mains to polluted rivers and questions over funding the ocean of necessary improvements, water’s reputation isn’t sparkling. Dig behind these headlines, however, and you’ll find an industry that’s embracing digital technologies — including AI — at a rate and with a radicalness that is entirely at odds with the sector’s outward facade.

The UK water industry’s reputation vs reality

‘Water touches everything we do and I think it’s undervalued’, says Jeannette Henderson, Principal, Ofwat Innovation Fund. ‘We can’t live without water… From data centres to swimming in rivers and the sea — it’s important and there are so many different facets. The more you dig into [the industry], the more you’ll find. It’s a fascinating area.’

Ofwat’s Innovation Fund: driving change in water utilities

Ofwat — the water industry’s economic regulator — founded its innovation fund in 2020 with an initial pot of £200 million. The aim was to encourage water companies to embrace technology through collaborating with academia, green NGOs, technology companies and engineering firms.

‘If you look back at the water sector, to 2019… We felt that water companies weren’t being collaborative and they weren’t being innovative’, Jeannette  explains. ‘They were being risk averse. Of course, there were good reasons for this — drinking water needs to be safe. But we wanted to stimulate innovation because we didn’t feel they could meet the big challenges by doing things the way things had always been done.’

The fund also aimed to tackle one of transformation and innovation’s biggest challenges: a business plan might run for five years, but an innovation, no matter how good, might take longer to create a return on investment.

Since its inception, the fund has supported 109 projects and has grown from its initial £200 million pot to £600 million. Examples of projects that received funding include using drones and satellites for near real-time water quality monitoring and pipe network imaging, AI for heat recovery from sewers, and 3D printing to create infrastructure with a lower carbon footprint. Other initiatives involve robotic slow sand filter cleaning, community-centric rainwater management, developing localised water supply schemes and converting sewage biogas into hydrogen.

‘Ultimately, the opportunities for change are limitless’, she says. ‘We have something like 300,000 miles of drinking water pipes in England and Wales — they would stretch from the Earth to the moon and beyond. There’s a huge infrastructure that needs monitoring and maintaining… What we’re seeing is that innovation is stimulating innovation.’

AI, robotics and smart water solutions

Looking beyond the moon and into seemingly sci-fi territory, the fund’s most futuristic project is leak detection using dark fibre. The project investigates using redundant fibre optic strands within the existing cable network for leak detection in water and wastewater systems. By firing lasers at intervals through cables that run adjacent to water mains, the team from Severn Trent, Welsh Water and Focus Sensors found it was possible to detect the noises associated with leaks. 
Testing revealed the system differentiates genuine leak signals from ambient noise, such as traffic and operational vibrations.

It’s actually sensitive enough to tell the difference between a human walking across a cable, and a fox’, Jeannette  says.

Early days for artificial intelligence

Clever sensing technology is one thing, but it is, of course, impossible to talk about transformation and change without mentioning AI. And here, Jeannette explains, it’s still early days for artificial intelligence. 

‘We’ve got a number of projects that use digital twins’, she says. ‘One of the projects is River Deep Mountain AI. They’re developing an open source model that informs how we tackle water body pollution. They’ve published some open data sets on GitHub and they’re encouraging people to test them and use them to develop deep learning systems… that will let us understand more complex systems like water bodies, so that we can make and accelerate change.’

Elsewhere, the imaginatively named AI and Sewer Defect Analysis project created a freely available benchmark dataset to accelerate the use of artificial intelligence in wastewater pipe inspections.

Water companies currently spend over £20 million annually on labour-intensive CCTV surveys, where engineers manually classify pipe defects. AI can automate this work and improve accuracy, but adoption has been slow because high quality training data is scarce.

To address this, several UK utilities shared coded CCTV footage with the Water Research Centre, which collated more than 725,000 images. After validation, over 27,000 clear, standardised images with defect metadata were released for public use. The dataset focuses on the most common defect types and follows strict criteria to ensure suitability for AI model training.

This first cross-sector, cloud-based resource enables AI providers to enhance their models, lowering costs and boosting productivity. At least four water companies are already using improved AI systems, and further investment could expand and maintain the dataset’s long-term benefits.

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

‘It’s an interesting project’, Jeannette says. ‘It sounds like it’s an AI tool. But it’s actually a data set. You have to remember, it’s not just the shiny stuff you can see. It’s actually the data that’s really important.’

Along with AI, robotics is also a technology that’s being explored as a means of boosting sustainability https://www.bcs.org/articles-opinion-and-research/data-centre-efficiency-and-the-impact-of-ai/ within the water industry. A project called Pipebot Patrol is creating an autonomous robot that lives inside sewers and is tasked with detecting early blockages and structural defects, giving maintenance teams time to prevent flooding and environmental harm.

Using robotics, acoustics and smart-sewer networks, the system integrates multiple sensors with a Raspberry Pi-based controller and microcontroller (MCU) for real time data capture, processing and wireless transmission. In partnership with EMS, Minicam and the University of Sheffield, the project is refining navigation, communication and planning field trials for a cost effective, long term monitoring solution.

The future of water: skills and careers in a digital sector

So, where can IT professionals fit into this world? What skills are most highly prized and which are in the highest demand? Jeannette  says: ‘Certainly cyber security. I was at a conference recently and there was a talk — there was a theme — around the need to grow skills in that area. The other thing I hear time and time again is the need for good data governance. We need to get the foundations right. It’s easy to develop a new AI or a new shiny approach… Some sort of model. But, you need to have good data. That can enable a lot more connectivity and interoperability. But, across the board there’s a demand for skills and this is a relatively new area for water companies. They’ve got lots of engineers but technology is a really big growth area for them.’ 
What does a digitally transformed water industry mean for customers?
Such an array of technologies and approaches being explored and deployed begs a final question: what will a digitally transformed water industry feel like from a customer’s perspective?

‘It’s an interesting question’, Jeannette  says. ‘But, I’d hope it would feel like a perfectly normal, everyday service. People wouldn’t notice the service or the technology. They’d turn their tap on and there’d be water there. There’d be no blocked pipes, no burst sewers. There wouldn’t be front page news. Water would go back to being boring.’