Why Predictive Modeling

Fast turnaround

Flexibile by design

Supported by experts

LCRI ready

17,000,000
+
Our AI model is built on the largest service line database in the world, over 17 million service lines.
Our solution scales to your system's needs, big or small.

Target your efforts. Lower costs. Make real progress.

Deadlines are coming fast. If your team suspects lead or GRR, 120Water's Predictive Modeling solution helps your team meet your compliance deadlines faster and at a lower cost than traditional verification methods.

Infographic titled 'Probable Lead Service Lines' with three categories showing counts: 14 with over 75%, 40 with 75-25%, and 73 with under 25%, above a partial map with colored location markers.

Prioritize the right locations

When funding is limited, every dig needs a reason. Focus on locations backed by data, not assumptions.

Construction worker holding a shovel standing next to a large excavator digging a trench in the ground.

Eliminate unnecessary digs

Nothing frustrates a crew like digging up clean pipe. Accurate predictions reduce low-value excavations so time and effort aren’t wasted.

User interface panel with expandable sections titled 'System-Owned' and 'Material'; under Material, details show Material: Non-Lead - PVC, Previously Lead: Unknown, Classification Basis: Modeling.

Confidence you can stand behind

You get predictions paired with real guidance from our team, so you know how to move forward and be in compliance with your state guidelines.

Here’s how it works

The 120Water team of experts will review your key data inputs and system specifications in great detail to train our AI model.

Webpage section showing details for 123 Chappell Rd including location coordinates, address verification scores, profile with LCRR Tier 5, sensitive population as No, unknown current sampling site and building plumbing container solder, and placeholder parcel information.

120Water runs the AI model and generates predictions to pinpoint where the highest lead and GRR risk lies.

Predictive modeling process infographic showing four steps: 01 Select a Customer, 02 Review Data, 03 Make Predictions, and 04 Review Results, connected by an ascending line graph.

You’ll review our predictions and conduct field verifications. This data then feeds into our platform and makes our model even smarter.

Plumber in an orange helmet and gloves repairing pipes in a construction site with exposed brick walls.

We confirm confidence levels meet requirements (95%+) or run more predictions if needed.

Chart showing over 95% confidence in meeting requirements, represented by a large green ring and two smaller rings for comparison.

We generate a deliverable designed to support your regulatory submission to your state.

Predictive report table showing two metrics: Negative Predictive Value (NPV) and Precision (Lead) with their definitions and placeholder values represented as X%.

View of El Paso city skyline at dusk with illuminated buildings and mountainous background, centered with City of El Paso logo featuring EPA TX and a star.

Predictive modeling in action

El Paso Water

El Paso Water, a large utility with approximately 240,000 service lines, entered 2025 with nearly 130,000 lines classified as unknown. After field-verifying about 60,000 lines in six months and encountering limitations due to inaccessible infrastructure, the utility turned to 120Water for predictive modeling. Using our model, more than 40,000 unknowns were classified as Non-Lead at an extremely low per-line cost.

This approach saved El Paso Water over $150,000, significantly reduced field work and customer notices, and helped cut the number of unknown service lines from roughly 130,000 to fewer than 13,000 in just one year.

130,000
Unknown service lines at the beginning of 2025
<13,000
Unknown service lines at the end of 2025
$150,o00+
Saved

Frequently Asked Questions

What is predictive modeling for water systems?

Our model analyzes dozens of variables including construction year, parcel data, housing characteristics, historical records, and known material types to create data-backed predictions for every address. Utilities can use these insights to prioritize field verification, reduce unnecessary digs, and accelerate their LCRI inventory updates.

How does 120Water’s predictive model work?

Notifications must be sent annually by July 1, unless state rules specify otherwise. With 120Water, we help automate your compliant letters and track delivery.

What data does my utility need to get started?

You don’t need perfect records. Our team works with whatever data you have: tap cards, GIS layers, billing data, parcel files, verification results, and replacement history. The model is designed to fill the gaps and enhance the information you already maintain, even if you have a large number of unknowns.

How can predictive modeling save time and budget for my water system?

Predictive modeling helps you avoid inspecting low-risk locations and instead focus verification efforts where lead is most likely to be found. That reduces field labor, cuts back on digs that aren’t needed, and helps you build a more accurate inventory with fewer resources. The result is faster progress with clearer justification for your decisions.

Is predictive modeling accepted by state primacy agencies?

Many states actively encourage data-driven verification strategies, and several allow predictive modeling as part of an approved statistical validation approach. Our team tailors every model to your state’s rules and provides documentation you can share with regulators, giving you confidence that your plan aligns with local requirements.

Can predictive modeling help beyond LCRI compliance?

Absolutely. Utilities use these insights for capital planning, prioritizing replacements, communicating more accurately with residents, and supporting future funding applications. Predictive modeling builds a stronger foundation for long-term water quality management, not just inventory completion.