With 4,700 field employees across 350 sites, Thames Water needed more than safety reports. They needed a system that captured what workers actually knew. Learning Teams gave them exactly that, and the results arrived within 90 days.
47%
More OLTs completed
3x
Frontline participation
64%
Actions closed on time
Zero
Missed actions in final 4 weeks
Thames Water serves more than 15 million customers across London and the Thames Valley. It runs 350 wastewater treatment sites, 100 clean water treatment works, and over 35,000km of pipe network.
Its operations team processes roughly 700,000 repair and maintenance jobs every year - managed by nearly 4,700 frontline employees working across multiple shifts, dozens of regions, and under constant regulatory pressure from Ofwat and the Environment Agency.
Thames Water was also deep into its most ambitious infrastructure programme in over 150 years, making operational clarity more urgent than ever.
Customers:
15 million+
Treatment sites:
350 wastewater, 100 clean water
Pipe network:
35,000km+
Frontline staff:
~4,700 field employees
Annual jobs:
~700,000 repair & maintenance
Pilot sites:
8 operational sites
Duration:
90 days
Thames Water had no shortage of safety processes. Toolbox talks happened. Risk assessments were filed. Incident reports went out on schedule. But the organisation kept running into the same problems at the same sites, with the same root causes buried underneath layers of documentation that nobody had time to read.
Four challenges were costing the business time, money, and trust:
Field engineers across Thames Water's treatment sites and pipe networks held deep, practical knowledge about how work actually happened. They knew which processes drifted under pressure and which equipment behaved differently in wet conditions. That knowledge never made it up the chain. There was no structured way to capture it, and no platform to act on it.
Thames Water's operations teams worked split shifts across geographically spread sites. Organising a face-to-face learning session meant pulling people off the job and across the city. Sessions were irregular. The teams that needed to be in the room the most were usually the ones who could not attend.
When learning sessions did happen, the outputs lived in spreadsheets or meeting notes. Actions were assigned verbally. There was no system to track whether anything actually got done. Managers followed up manually, which meant follow-through depended entirely on who was chasing and how much time they had.
Operational leaders at Thames Water received weekly and monthly reports. By the time those reports landed, the conditions that created the risk had already changed. Nobody could see which sites were quietly accumulating risk until something went wrong. Learning was happening after the fact, not before it.
Thames Water introduced Learning Teams to eight of its busiest operational sites during a 90-day pilot. The platform was designed to complement existing safety processes, not replace them. Facilitators were given guided onboarding, and within just two weeks of setup, they were running Operational Learning Teams (OLT) sessions.
The platform gave Thames Water three things in one place that had never been available before:
Phase 1 - Learn
Frontline engineers participated in structured OLT sessions, which were either held on-site or remotely via the platform. Each session lasted approximately two hours. Workers described how the work was actually done, not just how it was written in the procedure. The facilitators used the OLT flow built into the platform to keep the conversation organised and capture insights in real time. Teams from different shifts and sites were able to participate in the process without having to travel or disrupt routines.
Phase 2 - Soak
After each session, the platform stored all the collected information for a period of time so that participants could reflect on it without pressure. During this time, participants added additional observations, often not revealed in the initial session. This "soak period" brought to light important points that had previously been overlooked.
Phase 3 - Improve and Action
The teams returned for another short, focused session where the insights gained were translated into actionable improvements. Actions were assigned directly within the platform during the same session. Each action was assigned a responsible person, a due date, and a status that was updated in real time. Supervisors and operational leaders had a live view of every pending action across all eight sites, without having to send an email or chase a spreadsheet.
AI-powered theme detection
The same equipment appeared in four separate sessions across three different sites. The AI flagged it as a systemic pattern - leadership took action before it became an incident.
Within 90 days, Thames Water recorded clear and sustained improvements in all indicators that were being tracked during the pilot. These results were not due to additional staff or longer working hours.
The real difference was made by the platform, which removed the friction that had always limited the process of learning at scale.
47%
More OLTs completed vs. prior 90-day period
3x
Increase in frontline participation
64%
Of improvement actions closed on time
22
Cross-site risk themes identified by AI
2 wks
From onboarding to first completed OLT
Zero
Missed actions in the final 4 weeks of the pilot
Thousands of experienced engineers were going home every day with insights that were never stored within the organisation. Notes were made in meetings, turned into folders, and those folders were never opened. Learning Teams replaced this entire cycle with a structured platform that made learning consistent, trackable, and visible. The effects began to be seen within the first 30 days.
A clear process facilitators could follow. The 3-phase OLT model removed ambiguity and made every session productive, regardless of who was running it.
Remote capability removed the biggest barrier. Split-shift workers and geographically spread teams could participate consistently for the first time.
AI provided intelligence no spreadsheet could. Cross-site patterns emerged automatically, giving leadership the signal they needed before risks became incidents.
Learning from everyday work, not just accidents. The shift from reactive to proactive was visible within 30 days and compounded throughout the pilot.
Senior Operations Manager
Thames Water
“Our engineers always had the answers. We just never had a reliable way to get those answers into a room where decisions were made. Learning Teams changed that. Within the first month, we were surfacing issues that our incident reports had never captured.”
EHS Facilitator
Thames Water Field Operations
“We ran more OLTs in three months than we had managed in the previous year. The remote session capability made the biggest difference. Shift workers who had never been included before were now contributing every week.”
Learning Teams Operational Learning Software is built for operations teams that already work hard on safety
but still struggle to turn daily knowledge into lasting improvement.
Book a free 15-minute demo and see how your team could run its first OLT within two weeks.
Empowering Insights, Driving Excellence: Transforming Work with Operational Learning.
Copyright © 2026 Learningteams™. All Rights Reserved.
