AI Project of the Year
Balibot
What is it?
Balibot AI Knowledge Twin is an AI assistant for asset management, preventive maintenance and leak management for Balibago Waterworks in the Philippines.
Who is involved?
Balibago Waterworks is the client and TeamSolve, a Singapore-based start-up, developed the Knowledge Twin.
What makes it special?
- Balibot has enabled Balibago Waterworks to effortlessly transition from manual data gathering and ad-hoc scheduling to instant field-data capture and insight generation, helping prioritise maintenance with AI-guided workflows. Automated reporting centralises knowledge and speeds up crew mobilisation, with field crews reporting up to 50% time savings.
- AI has proved game-changing for Balibago’s non-revenue water strategy, where recovered revenue is precious in an underfunded sector. Balibago had assumed mainline pipes were the problem, but leak survey reports from Balibot showed service connections were a larger issue, shifting strategy to replacing old service connections. Balibot also identified the pipe material driving the most leaks and recommended how to prevent future leaks, enabling a more proactive stance.
- Scaling AI solutions is difficult without organisational buy-in. Crucial to the success of the project was the focus on adoption: continuous training of around 100 plumbers and pump operators and strong executive buy-in were paramount to the utility embracing the mindset of digital transformation. Five municipal franchises and two water treatment plants are now being optimised thanks to Balibot.
SEDIF digital asset modelling
What is it?
AI-driven 3D digital asset modelling for France’s largest drinking water authority, the Syndicat des Eaux d’Île-de-France (SEDIF), supplying 4 million people across 133 municipalities. The asset scope is three water treatment plants, 91 reservoirs, 43 pumping and 45 chlorination stations, with 60,000 pieces of equipment modelled.
Who is involved?
Veolia Franciliane is the delegated operator of SEDIF and deployed the AI-based solution at full operational scale. The solution was supplied by French start-up Samp.
What makes it special?
- Digitising tens of thousands of assets demands quality and speed. Starting from 3D scans, AI automatically recognises and structures equipment directly in the real environment, delivering accurate, asset-based 3D reality models within days to support an accelerated timeline. One site can be fully mapped in a day, a step change in efficiency for a project of this magnitude.
- 3D scans link easily to operational and maintenance data in the utility’s computerised maintenance management system, replacing fragmented documentation approaches with a shared, reliable operational reality across SEDIF’s production sites and slashing the need for costly site visits.
- The project shows AI can be deployed at true industrial scale, with Veolia brilliantly bridging the gap between AI and field-level maintenance to ensure optimised maintenance and operational excellence for critical drinking water infrastructure serving four million people.
Shenzhen AI Leakage Control
What is it?
The Shenzhen AI Leakage Control Project is a proactive, mega-city water loss management system serving 18 million residents across a 20,000km distribution network.
Who is involved?
Shenzhen Water and Environment Group is the client and developer, partnering with the Water Authority of Shenzhen Municipality, Alibaba Cloud, and Baidu.
What makes it special?
- This project fundamentally shifts mega-city water management, transitioning Shenzhen from reactive reporting to proactive, AI-optimised intervention for infrastructure upgrades. Powered by over 3.5 million smart meters alongside a city-wide array of flow, pressure, and acoustic sensors, the AI ingests real-time data to pinpoint leaks with a scale and precision unattainable by human inspection.
- The system’s true power lies in integrating machine learning with China’s largest urban hydraulic model. With over 1.6 million simulation nodes, the AI draws on 300,000 real-time monitoring points to cross-validate live data against predicted outcomes. This revolutionised emergency response, slashing the time to locate and repair leaks across thousands of DMA zones from 30 days to 10.
- Ultimately, this integration delivered a critical breakthrough, with Shenzhen’s water loss rate dropping to 4.4% in 2025 after stalling at 4.6% for three years. Achieving the lowest rate among China’s first-tier cities, it sets a world-class benchmark translating to massive gains: conserving 3.5 million cubic metres of water a year.
SWW Sewer Optimisation
What is it?
UK utility South West Water (SWW) is turning big data from its Event Duration Monitoring of sewer overflows into actionable insights, using AI to reduce sewer spills and ensure targeted environmental action, all while avoiding false positives.
Who is involved?
Metasphere – part of Grundfos – provided the MAP analytics platform for SWW and Grundfos provided predictive AI capabilities.
What makes it special?
- SWW previously faced 95% non-genuine alerts when monitoring sewer overflows, creating alarm fatigue that risked burying critical failures while teams wasted time investigating non-existent problems. Grundfos integrated datasets like real-time rainfall and ground saturation and used 12-hour predictive sewer-level AI to filter up to 9,500 phantom alerts per month so SWW could focus on the right issues.
- With 95% alert accuracy, the shift from reactive monitoring to AI-driven intelligence turned huge datasets into a precision tool, proactively clearing 240 confirmed blockages before they could become spills. The result was a 50% year-on-year reduction in storm spills and pollution incidents and the removal of over 7,000 avoidable blockages, underscoring the need for targeted intervention.
- A sophisticated “management-by-exception” approach restored systemic integrity and renewed confidence in SWW’s digital infrastructure. Grundfos’ scalable, low-cost AI model helps utilities turn mountains of data into usable insights, invaluable for achieving swimmable rivers in the UK and beyond.
Tampa Bay Early Detection Water Quality Monitoring System
What is it?
An intelligent platform capable of predicting turbidity levels, salinity spikes and red tide episodes, among other critical parameters, through monitoring data in the bay and open water prior to the intake of Tampa Bay Seawater Desalination Facility.
Who is involved?
Acciona, in collaboration with its technological partner SkyTL. Tampa Bay Water owns the desalination facility.
What makes it special?
- The project analyses real-time data from both the bay and open waters, enabling the prediction and notification of incidents at least two hours before the instruments installed at the plant. This lead time allows operators to take proactive measures and avoid facility shutdown, maintaining efficient, safe, and reliable operation of the Tampa Bay desalination plant.
- Machine learning analyses data from a dizzying array of information sources, such as weather, runoff, bay discharges as well as fixed cameras and drones. Acciona and SkyTL’s expert combination of ingesting diverse data and predictive algorithms provides a comprehensive view that enhances operational reliability and strengthens water security for the region.
- The system also optimises chemical usage, extends equipment lifespan and reduces emergency repairs. With desalination becoming an increasingly important water security option, the platform is a huge step towards ensuring such facilities can continue to operate in the face of ever extreme climate events.