The UK Water industry is facing direct pressure to improve its services. After the closure of Asset Management Period (AMP) 7, leakage remains high, dirty water discharge incidents continue to draw great scrutiny, and existing infrastructure continues to age. AMP8 sets clear requirements to change this landscape through complete digitalization and trusted and accurate data-driven operations.
In June 2026, Ofwat introduced the AI Adoption Plan for the water sector to reinforce this directive. AI has already seen its position change from an experimental phase to a core operational capability. It is now a part of our critical national infrastructure.
Now more than ever, the quality of our data is crucial because of the growing importance of digitization and artificial intelligence in our systems.
The impact of regulation on operations
It's been a year since the AMP8 cycle started, and water companies are under pressure to speed up and act sooner. The rules in place are meant to not only meet the AMP8 goals but also boost the achievements of the previous AMP7 objectives. Looking back at the environmental performance ratings in the Environment Agency's yearly report, it's clear that most companies are still not doing well enough, as the report states, "the majority of companies continue to underperform." There's still a lot of work to be done to improve their environmental track record.
By using AI, water companies can better understand how their assets are working and spot any unusual activity in the data from their networks. This helps them predict how much water will be needed in the future. These things are closely tied to the goals that regulators set. If companies can detect leaks early, they can reduce the amount of water that is lost. For example, Machine Learning, a type of artificial intelligence, can help predict when and where leaks are likely to happen, which also reduces strain on sewer networks and can lower the frequency of storm overflow discharges. It can do this by showing where the network is under stress. Additionally, by performing predictive maintenance, identifying which assets are likely to fail before they actually do, companies can make sure their systems are reliable. This is important for providing a steady supply of clean water.
The regulators’ position is therefore clear; AI is not to be used purely for efficiency, it must deliver measurable improvements to customers, and soon.
The criticality of existing data and breaking down silos
Data is the single biggest constraint facing water companies from successfully achieving these objectives. For example, telemetry data can differ from one part of the network to another, historical records contain gaps, and asset registers do not always reflect the current state. Most high-value data reside within operational systems. Systems that have been designed to operate reliably, and not necessarily to feed analytics into decision engines. Sensors measure flow, pressure and quality, SCADA platforms manage control processes. Processes exist, large quantities of data exist, but the effort to prepare the data for wider consumption and valuable insight takes considerable time and effort.
To really make a difference, AI needs to bring together two kinds of data: the day-to-day operational data and the bigger picture enterprise business data. It's all about understanding how everything fits together in the organization. For AI to truly deliver what the industry needs - and what regulators and customers expect – it's essential that operational technology (OT) and information technology (IT) teams work hand-in-hand. By combining their strengths, they can give AI the best chance to succeed and make a real impact.
Data accuracy and the importance of assurance
When it comes to AI models, getting the data just right is crucial. Even tiny mistakes can throw off the whole decision-making process downstream. For instance, faulty sensor readings or incomplete maintenance records can trigger unnecessary inspections and maintenance activities, diverting resources away from genuine operational priorities. These issues are a direct result of fragmented and outdated data not in the AI models themselves. Older systems and processes could absorb these inconsistencies, human error for example can be lost in aggregated data sets. AI operates at a different level. It needs consistency of data across time, location, and format.
Water companies are now expected to manage data with the same discipline applied to physical assets. This requires continuous monitoring of data quality, clear standards for collection and storage, and assurance processes that identify issues at source. Consistent controls, data lineage tracking and quality validation help ensure AI models are built on reliable information. Without these foundations, confidence in AI outputs remains limited, slowing adoption and reducing the value of investment in digital transformation.
The power of data governance
To meet AMP8 objectives and realize the benefits of AI, organizations need clear governance through ownership of datasets, defined accountability for data quality, and consistent processes for updating, managing and sharing information across teams. The implications of weak governance are familiar across many organizations. Different teams often maintain their own datasets to support local requirements and over time, multiple versions of the same information emerge and in turn enterprise views across operational reports produce conflicting results. AI models trained on different sources generate inconsistent recommendations. As confidence in the data declines, decision-making slows and reliance on manual validation increases.
Good management helps solve problems by giving everyone in the company the same accurate information. When you connect the day-to-day data from your systems and equipment with your planning tools, reports to the government, and business processes, things get easier. Your teams won't have to waste time checking if the information is correct, and they can focus on acting instead. As AI becomes more embedded in decision-making, trusted data enables confidence in its recommendations.
Overcoming the challenges for AMP8 success
While every water company is at a different stage of its digital transformation journey, the same challenges appear consistently. Legacy infrastructure, data quality and a focus on system safety and resilience rather than speed of change are all individual contributors to poor implementations of AI.
To overcome these challenges, it's essential to focus on good design, making sure data is accurate and reliable, and having strong governance in place. By doing this, organizations can turn their digital plans into real successes. Companies that get all three of these things right are more likely to achieve their goals and make the most of artificial intelligence. This approach helps ensure that digital ambitions lead to positive results, which is crucial for meeting objectives and realizing the full potential of AI.