Book A Call
Case Study · Water & Wastewater Utilities

How Irvine Ranch Water District turned maintenance data into a decision system

Built around the business drivers, not the data dump. FreshBI turned a water district’s maintenance data into a decision system, where every measure ties to a business driver, a target and an owner.

Talk to us about your business See the five drivers
5
business drivers, each with its own measures, a target and an owner
12+
linked views, from the whole district down to a single asset
~2%
share of corrective labor hours going to emergency work, tracked every week
<20%
district-wide overtime held below target, with hotspots flagged by work group
  1. 1 Business first Start from the business drivers, not the data dump.
  2. 2 Ontology first Every measure gets a definition, a target and an owner in one governed model.
  3. 3 AI ready Governed, secure and ready for plain-language questions and AI agents.

About Irvine Ranch Water District

Irvine Ranch Water District (IRWD) serves a fast-growing region of Southern California with drinking water, wastewater collection and treatment, and an internationally recognized water-recycling program.

That means maintenance at scale: thousands of work orders a year across plants, pump stations, wells and several specialist crews.

Business first

Plenty of data, no line of sight

The district’s asset management system captured every work order, labor entry and status change. But volume isn’t insight.

Activity, not performance

Reports showed orders opened and closed. They didn’t show whether the maintenance program was reliable, under control and sustainable.

Are we doing the right work, or just the urgent work?

Answered by work control: how much labor is planned vs. lost to reactive work.

Where is work getting stuck, and why?

Answered by throughput and supply readiness: how fast work finishes, and how long crews wait on parts and materials.

Are our crews and assets set up to stay ahead of failure?

Answered by reliability and labor: whether preventive work is on schedule, and where overtime and repeat problems concentrate.

Ontology first

Drivers first, dashboards second

FreshBI didn’t start with charts. Working with operations and maintenance leadership, we pinned down the business drivers that actually decide maintenance performance at a utility, then defined each measure once.

Every measure passes one test

If this number moves, who acts, and what do they do? Measures that failed the test were cut.

A definition, a target and an owner

The measures that remained each got a clear definition, a target and an owner, so every trend shows its goal line.

One version of the truth

A single governed model with certified definitions for every measure, and every page explains exactly how its measures are calculated.

The solution

Five drivers, one operating picture

FreshBI delivered a governed analytics model over the district’s enterprise asset management data, as a guided, multi-page Power BI performance suite with 12+ linked views from the whole district down to a single asset.

01 · The five business drivers

What decides maintenance performance at a utility

Each driver has its own measures, a target and an owner.

  1. 1

    Reliability discipline

    Is preventive work on schedule, and catching problems early?

  2. 2

    Work control

    How much labor is planned vs. lost to reactive work?

  3. 3

    Throughput & backlog

    How fast is work finished, and how much is aging?

  4. 4

    Supply readiness

    How long are crews kept waiting on parts and materials?

  5. 5

    Labor sustainability

    Is overtime on target? Which assets eat the most labor?

02 · From driver to decision

Each driver links to the measures that prove it and the action they trigger

That link is why the dashboards drive weekly planning.

  • Reliability: PM compliance against target by crew, and corrective work generated from PM inspections. It drives rebalanced PM schedules and checks that inspections are finding real issues.
  • Work control: planned vs. unplanned labor by week, and the emergency (top-priority) share. It shifts capacity toward planned work and prompts investigation of emergency spikes.
  • Throughput: backlog in weeks of work vs. a healthy band, an aging heat map, and mean time to complete. It drives targeted backlog clean-ups and staffing for the crews with the longest queues.
  • Supply readiness: orders waiting on material, with mean and longest wait by crew and age. It escalates stalled procurement and pre-stages parts for planned jobs.
  • Labor: overtime % vs. target, top assets by labor hours, and the location hierarchy. It finds overtime hotspots and flags repeat-problem assets for repair-or-replace review.
03 · Designed for action, not admiration

Dashboards that turn into work lists

The suite is built to be used in weekly planning, not just looked at.

  • A target on every trend. Each chart shows the goal line, so performance reads at a glance.
  • Plain-language status. Headline tiles carry On Track, Monitor and Critical flags, not just numbers.
  • District to crew to asset. One click drills from the district view to a work group, a facility or a single asset.
  • Insight that becomes a work list. The backlog clean-up view turns aging orders into a prioritized list, with each order’s history alongside.
  • Definitions built in. Every page explains exactly how its measures are calculated.

Governed, secure, trusted

One version of the truth

A single governed model, with certified definitions for every measure.

Least-privilege access

Read-only data connections and role-based security.

Data stays home

Data stays inside the district’s own environment and controls.

Quality you can see

Monitored refreshes, documented lineage and reconciliation checks.

The results

What changed for IRWD

  • Backlog health. Open work is measured in weeks of crew capacity against a healthy band, so leaders see when it outruns the team.
  • Aging work. Every order past 90 days is sorted into age buckets by crew, then turned into a clean-up list someone owns.
  • Emergency share. About 2% of corrective labor goes to emergency work. It’s checked weekly, so any spike gets looked into.
  • Overtime. District-wide overtime runs below the 20% target, and any crew over target is flagged.

Gaps like PM compliance by crew and repeat-labor assets are now visible and owned, which is the first step to fixing them. Results come from the client’s own dashboards; operational detail is summarized to protect client confidentiality.

AI ready

Built AI-ready from day one

The same principles that make these dashboards trustworthy (one governed model, certified definitions, clean history and secure access) are exactly what AI needs to work. That leaves the district positioned for the next step, all grounded in data it already trusts.

Plain-language questions

Ask the governed model questions in everyday language and get answers built on certified definitions.

Failure-risk prediction

Predict which assets are most likely to fail, using the clean maintenance history already in the model.

Agents that draft work plans

AI agents that draft work plans, grounded in data the district already trusts.

The FreshBI method: decisions first, AI-ready by design. Read more in AI-Ready Data, explore our Power BI dashboard services, or see our data platform architecture approach.

Maintenance analytics and AI readiness: common questions

How do you design maintenance analytics for a water utility?

Start with the business drivers, not the data. At Irvine Ranch Water District, FreshBI worked with operations and maintenance leadership to pin down five drivers: reliability discipline, work control, throughput and backlog, supply readiness, and labor sustainability. Every measure had to pass one test: if this number moves, who acts, and what do they do? Measures that failed were cut, and the rest got a definition, a target and an owner. See our Power BI dashboard services.

What does "drivers first, dashboards second" mean?

It means the dashboard is the last step, not the first. The drivers and the decisions they support are defined first, then each measure is tied to an action. That is why the IRWD dashboards drive weekly planning instead of just reporting activity.

How is utility data kept secure in a Power BI performance suite?

The solution uses read-only data connections and role-based security, and the data stays inside the district’s own environment and controls. Refreshes are monitored, lineage is documented, and reconciliation checks keep the numbers trustworthy.

What makes analytics AI-ready?

One governed model, certified definitions for every measure, clean history and secure access. Those are the same things AI needs to give reliable answers, which is why IRWD is positioned for plain-language questions, failure-risk prediction and agents that draft work plans. See AI-Ready Data.

How long does a project like this take?

FreshBI typically delivers a first production solution in as little as three weeks. See pricing.

More case studies
Logistics Johanson Transportation 4 AI products in production on one governed platform Read the case study → Logistics Chicago Intermodal <1 min from tender acceptance to a complete order in the TMS Read the case study → Forest products COFI 20 days from manual reports to real-time dashboards Read the case study → Food & CPG Nature’s Path 1 shared dataset behind every innovation decision Read the case study → All case studiesBrowse the full collection →

Ready to make your data decision-ready and AI-ready?

Governed analytics and AI integration for utilities, public agencies and the enterprise.

Book A Call