FIDO AI

Learning Objectives

• Describe what FIDO AI does and why leak detection matters for water utilities
• Explain how machine learning on acoustic data estimates not just the presence but the size of a leak
• Identify the concept of non-revenue water and why prioritizing large losses matters

What Is FIDO AI?

FIDO AI is an artificial intelligence system for finding leaks in water distribution networks. Built by FIDO Tech, an Oxford, United Kingdom company founded in 2014 and backed by SKion Water, the platform analyzes acoustic and sensor data from water pipes to detect leaks — and, distinctively, to estimate how big each leak is. Most leak-detection tools tell a utility that a leak exists somewhere; FIDO goes further by sizing the loss, so operators can direct crews to the biggest leaks first.

That sizing capability directly attacks one of the water sector's most expensive problems: treated water that leaks out of the network before it ever reaches a paying customer. FIDO reports strong accuracy across pipe materials, and its FIDOSmart app extends the same AI leak analysis into a low-cost smartphone workflow, making the technology accessible far beyond large utilities. The company has collaborated with operators including United Utilities.

What FIDO AI Does

• Acoustic leak detection — analyzes sound and sensor signatures from pipes to identify leaks in the network
• Leak sizing — estimates the size of each leak so utilities can prioritize the biggest losses first
• Material-agnostic accuracy — works across a range of pipe materials rather than a single type
• FIDOSmart app — brings AI leak analysis into a low-cost smartphone workflow for field teams
• Non-revenue water reduction — focuses effort where it recovers the most lost water and cost

How AI Is Applied

FIDO's AI is genuine machine learning, and it is the core of the product. The system is trained to interpret acoustic and sensor data from pipes, learning to distinguish the signal of a leak from background noise across different pipe materials and conditions. Crucially, the models do not just classify leak or no-leak — they estimate the magnitude of the loss, which requires the AI to relate subtle characteristics of the acoustic signature to the physical size of a leak.

That size estimate is what turns detection into prioritization. A utility with limited crews cannot chase every leak at once, so knowing which leaks are draining the most water lets them recover the largest losses first. By packaging the same models into the FIDOSmart smartphone workflow, FIDO makes advanced leak analysis usable without expensive specialist equipment — a meaningful step for smaller utilities and for water-scarce regions.

Who Uses FIDO AI

FIDO AI is used by water utilities and network operators working to reduce leakage and non-revenue water. It serves large utilities running network-wide leak programs as well as smaller operators and field teams who can adopt the AI through the low-cost FIDOSmart app. Water-scarce regions, where every recovered liter counts, are a natural fit.

Pricing

FIDO AI is sold to utilities on an enterprise, quote-based model, with the lower-cost FIDOSmart app broadening access for smaller teams. Pricing depends on network size, deployment scope, and the mix of sensors and app-based analysis. Utilities contact FIDO Tech directly for a tailored quote.

Company Details

Strengths

• Leak sizing — estimates how big each leak is, enabling utilities to prioritize the largest losses first
• Genuine machine learning — real ML on acoustic and sensor data, not just thresholding or manual review
• Material-agnostic — maintains accuracy across different pipe materials
• Accessible via FIDOSmart — a low-cost smartphone workflow puts AI leak analysis within reach of smaller utilities
• High-impact target — directly reduces non-revenue water, a costly and wasteful problem

Limitations and Considerations

• Data quality dependence — acoustic detection accuracy depends on sensor coverage and signal conditions in the field
• Detection, not repair — FIDO finds and sizes leaks; utilities still need crews and processes to fix them
• Network context — best results come when the tool is integrated into a broader leakage-management program
• Estimates remain estimates — leak-size figures guide prioritization but are approximations of physical losses

Key Takeaways

• FIDO AI applies genuine machine learning to acoustic and sensor data to detect water-network leaks and estimate their size
• Sizing leaks lets utilities prioritize the biggest losses first, directly attacking non-revenue water
• The FIDOSmart app extends the same AI into a low-cost smartphone workflow, widening access beyond large utilities
• Best for water utilities and field teams focused on reducing leakage and recovering lost, treated water