CSO LAUNCH PARTNER PROGRAMME · 2026

Build the next generation of CSO monitoring — with us.

We're inviting a small number of progressive water utilities to co-develop continuous, vision-AI-backed monitoring for combined sewer overflows. Our platform is proven on distributed water assets at VA SYD and EMASESA. The CSO-specific configuration — wired wake-on-trigger tags, ATEX-rated housing, regulator-grade audit trail — is what we're building together.

Limited launch-partner intake · 2026

CSO monitoring infrastructure
66 %
fewer site visits at VA SYD
15 days
to actionable data at EMASESA
3–5 yrs
battery life per W-Sensor
The launch partner programme

A partnership, not a procurement.

Most utility innovation pitches sell finished products. This isn't one. We're bringing a proven platform and inviting a small number of utilities to shape the CSO-specific configuration with us — in exchange for first-mover positioning, programme-level pricing, and direct influence over the product roadmap.

PROVEN TODAY

What we bring on day one

  • The W-Sensor platform — battery-powered, embedded cellular, edge AI inference. Deployed at scale at VA SYD and in pilot at EMASESA.
  • Mimir cloud platform — fleet management, dashboards, alerts, image storage, exports, API integrations. In production today.
  • Vision AI training pipeline — a handful of site images is enough to produce a working severity model, as proven at EMASESA in 15 days.
  • Edge AI processing — 0–100 severity scoring on-device, anomaly flagging, time-stamped image series.
IN ACTIVE DEVELOPMENT

What we co-develop with you

  • Wired wake-on-trigger tags — instant wake the moment a float or threshold is crossed. Image capture timed to the event itself, not the next scheduled poll.
  • ATEX-rated housing — sensor and tag enclosures certified for hazardous-area Zone 2 deployments, designed for sustained H₂S and methane exposure.
  • CSO-specific AI models — trained jointly on your chamber types, debris patterns, flow regimes, and the specific events that matter for your reporting workflow.
  • Regulator-aligned export formats — visual audit trails packaged to complement EDM data in your jurisdiction's reporting cadence.
What launch partners get
01
First-mover positioning
Public reference for a category that's about to matter to every regulated utility. Joint case study, conference profile, peer credibility.
02
Direct product influence
Your operational requirements drive the spec. The CSO sensor housing, tag form factor, and AI model architecture are shaped by your network.
03
Programme-level pricing
Pilot-phase commercials reflect the partnership. Locked-in rates for production rollout if the pilot validates.
04
Embedded engineering
Direct access to our hardware, AI, and platform teams during the pilot. Issues get resolved fast because the people who can fix them are the people answering the call.

Why CSO monitoring needs more

EDM coverage was the easy part. Defending it is harder.

With monitors on virtually every storm overflow across the UK, EU, and US regulated markets, the question isn't whether utilities have data — it's whether the data holds up under public, regulator, and legal scrutiny. The water companies themselves have admitted what EDM alone can't do:

The data isn't always accurate. The monitors are sensitive and can play up. Even the movement of a weed growing in front of the monitor could set it off, indicating that the overflow is active when it isn't.

Thames Water · public guidance on storm discharge and event duration monitoring

No monitoring equipment is ever 100% accurate and free of erroneous signals. Discharge signals can be received as a result of company or contractor actions — routine maintenance, chamber entry or other intervention.

CIWEM Urban Drainage Group · EDM Good Practice Guide

False positives don't just inflate spill counts. They invite challenges, complicate enforcement, and make trend data noisier than it should be — right when regulators are using that data to drive billions in capital investment decisions. The platform we're developing with launch partners closes that gap.

EDM-only vs EDM + vision AI

Both signals, one platform.

The CSO programme doesn't replace your EDM — it pairs continuous camera-based proof alongside it. Wired wake-on-trigger tags mean image capture is timed to the actual event start, not the next polling window. The result: graduated severity, false-positive filtering, and a visual audit trail every spill event can be defended with.

EDM alone

Binary, indirect, contestable

Level sensor reaches a threshold; an event begins. The monitor tells you when, but not what.

  • Binary on/off signal — no severity, no context
  • False positives from debris, growth, chamber entry, sensor drift
  • No visual record of what was actually happening
  • Manual data correction often required before reporting
  • Field crews dispatched to verify ambiguous signals
  • Hard to defend trigger validity under regulatory or media scrutiny

EDM + Waltero CSO programme

Graduated, continuous, evidenced

Same EDM signal paired with wired wake-on-trigger image capture, AI scoring, and a permanent visual record.

  • 0–100 severity score, mapped to four operational bands
  • Wake-on-trigger captures the moment an event begins
  • Visual confirmation filters out false positives at source
  • Time-stamped image series for every event — the audit trail
  • Cleaning effectiveness verified remotely, no return visit needed
  • Defensible evidence for regulators, public, and litigation

How it works

One device. From asset to action.

The W-Sensor is a single battery-powered device that captures, analyses, and reports. Wired tags wake the device the moment a float or threshold is crossed — ensuring image capture is timed to the actual event, not whatever the next scheduled poll happened to be.

Waltero CSO monitoring architecture: from CSO chamber sensors through edge AI and cellular network to the Mimir cloud platform.

FIELD
In the CSO chamber
EDGE
On the device
NETWORK
Cellular uplink
PLATFORM
Cloud + dashboards

W-SENSOR

Camera + AI device

Battery-powered, 3–5 yrs

WIRED TAGS · WAKE-ON-TRIGGER

  • Event trigger — Float / threshold
  • Level / depth — Continuous on wake

wired · instant

image

Edge AI

processing on the sensor

  • Image capture
  • Vision inference
  • Score 0–100
  • Anomaly flag

no · low · med · severe

Trained on a handful of site-specific images

scored data

Cellular uplink

no gateway needed

  • NB-IoT / LTE-M
  • Embedded SIM
  • Encrypted in transit

Mimir

Waltero cloud platform

  • Live dashboards
  • Threshold alerts
  • Visual audit trail
  • Regulatory exports
  • API integrations

Field crews, regulators, and SCADA — one platform

One device · battery-powered · install in hours · operational data in days

The CSO-specific wired tag variant and ATEX-rated sensor housing are in active development — launch partners co-define the final spec.

Adjacent proof · Case Study 01
🇸🇪 VA SYD · Sweden

66% fewer site visits across distributed water-inlet infrastructure.

VA SYD is a regional water and wastewater utility serving multiple municipalities across southern Sweden. Field-crew time dominated their operating budget for distributed water-inlet inspection — with most scheduled visits revealing no required action.

After deploying W-Sensors across scheduled water-inlet facilities, two-thirds of routine inspections were eliminated. Continuous camera-based monitoring through the Mimir platform replaced in-person rounds. They're now extending into pump vibration sensing and infrared cameras — the same platform expanding into adjacent monitoring problems.

−66%
site visits eliminated
Multi-municipality
distributed network
Now expanding
to pumps + infrared

VA SYD is not a CSO deployment. It's proof the platform delivers when applied to the same operational pattern — distributed assets, scheduled inspection, debris and blockage detection. CSO is the next chapter we're writing.

Vision AI monitoring of raw water intake
Adjacent proof · Case Study 02
Mimir platform showing how stormwater blockage is tracked
🇪🇸 EMASESA · Spain

Spain's 4th-largest utility ran a 15-day pilot — and captured the full blockage lifecycle remotely.

EMASESA serves 1.2 million consumers across the Seville metropolitan area, operating thousands of distributed water assets. Stormwater grates regularly block with debris — wet wipes, organic matter, litter — with significant flooding consequences when blockages go undetected.

Two locations were chosen to stress-test the system. At Felipe González, three W-Sensors covered back-flow plus grate surveillance with both narrow- and wide-angle lenses. At Las Aseñas, sensors monitored outlet collection bags and the back grid — capturing the complete operational lifecycle of a blockage.

AI BLOCKAGE SCORING · 0–100 SCORE → 4 OPERATIONAL BANDS
No blockage
0–10
Low
11–40
Medium
41–70
Severe
71–100

Trained on a handful of site-specific images — no labelled datasets, no ML team, no months of tuning. Within 15 days the model was producing continuous severity scores EMASESA could act on.

Severe blockage detected → Cleaning performed → Medium blockage → No blockage confirmed

The operational proof point: EMASESA's team confirmed cleaning effectiveness remotely, without dispatching a crew for verification. This is the same audit-trail capability launch partners get for CSO events — only with wired wake-on-trigger image capture timed to the discharge itself.

15 days
to actionable data
1.2M
consumers served
2 sites
full lifecycle captured
Pilot economics

Model your scenario directionally.

Before the partnership conversation, this is a useful starting point: estimate the operational impact for a pilot scoped around your network and inspection regime. Pricing reflects volume tiers automatically and is clearly labelled as pilot-stage indicative.

CSO Launch Programme · Pilot Economics

Model your pilot scenario

Estimate the operational impact for a pilot scoped around your network. The numbers below are directional inputs to the partnership conversation — not a quote.

Open the model →

Pilot Economics Model · CSO Launch Programme

A directional model for the partnership conversation. Adjust the inputs to reflect your network and inspection regime; the figures will inform pilot scoping, not a binding quote. i Pilot-stage indicative pricing. Final commercials are agreed jointly during programme onboarding.

Configure Pilot Scenario

Your Network
const [points, setPoints] = useState(50);
  const [installCost, setInstallCost] = useState(600);
  const [inspSimple, setInspSimple] = useState(300);
  const [inspComplex, setInspComplex] = useState(600);
  const [inspPerYear, setInspPerYear] = useState(15);
  const [avoidedPct, setAvoidedPct] = useState(50);
  const [tagQty, setTagQty] = useState({ event_trigger: 1, level_depth: 0 });
Monitoring pointsvalue="50"
min="10" max="2000" step="5" value="50"
Installation cost/point$value="600"
min="0" max="2000" step="10" value="600"
Inspection cost (simple)$value="300"
min="0" max="1500" step="10" value="300"
Inspection cost (complex)$value="600"
min="0" max="3000" step="25" value="600"
Inspections/point/yearvalue="15"
min="1" max="52" step="1" value="15"
Avoided with continuous monitoringvalue="50"%
min="0" max="100" step="1" value="50"
Wired Tag Configuration
Wired event trigger

1 per sensor

Decrease Wired event trigger quantity −

Increase Wired event trigger quantity +

Float or threshold sensor. Wakes the W-Sensor instantly when an event begins — image capture is timed to the event itself, not the next scheduled poll.
$80*/tag at 10–99 pts
Wired level / depth
Continuous level reading once triggered. Captures fill rate, peak depth, and event duration with high resolution while the chamber is active.
$350*/tag at 10–99 pts
Indicative Pricing — Volume Based
const SENSOR_TIERS = [
  { min: 10,   max: 99,    label: "10–99",   sensor: 425, saas: 12 },
  { min: 100,  max: 249,   label: "100–249", sensor: 400, saas: 11 },
  { min: 250,  max: 499,   label: "250–499", sensor: 365, saas: 10 },
  { min: 500,  max: 999,   label: "500–999", sensor: 330, saas:  8 },
  { min: 1000, max: 99999, label: "1000+",   sensor: 290, saas:  7 },
];
 
const TAG_TIERS = {
  event_trigger: [
    { min: 10,   max: 99,    price: 80 },
    { min: 100,  max: 249,   price: 75 },
    { min: 250,  max: 499,   price: 70 },
    { min: 500,  max: 999,   price: 60 },
    { min: 1000, max: 99999, price: 55 },
  ],
  level_depth: [
    { min: 10,   max: 99,    price: 350 },
    { min: 100,  max: 249,   price: 330 },
    { min: 250,  max: 499,   price: 300 },
    { min: 500,  max: 999,   price: 270 },
    { min: 1000, max: 99999, price: 240 },
  ],
};
function sensorTierFor(points) {
  for (const t of SENSOR_TIERS) { if (points >= t.min && points <= t.max) return t; }
  return SENSOR_TIERS[SENSOR_TIERS.length - 1];
}
function tagTierFor(tagKey, points) {
  const list = TAG_TIERS[tagKey] || TAG_TIERS.event_trigger;
  for (const t of list) { if (points >= t.min && points <= t.max) return t; }
  return list[list.length - 1];
}
const sensorTier = sensorTierFor(points);
const sensorCost = sensorTier.sensor;
  const saasPerPointMonth = sensorTier.saas;
 
  const tagsBom = TAG_OPTIONS.map(t => {
    const qty = tagQty[t.key] || 0;
    const tier = tagTierFor(t.key, points);
    return Object.assign({}, t, { qty, price: tier.price, perSensorCost: qty * tier.price });
  });
  const selectedTags = tagsBom.filter(t => t.qty > 0);
  const tagsCostPerSensor = selectedTags.reduce((s, t) => s + t.perSensorCost, 0);
 
  const hardwareCost = points * sensorCost;
  const installation = points * installCost;
  const annualSaaS = points * saasPerPointMonth * 12;
  const tagsTotal = points * tagsCostPerSensor;
 
  const totalInspectionsPerYear = points * inspPerYear;
  const inspectionsAvoided = totalInspectionsPerYear * (avoidedPct / 100);
  const avgInspectionCost = (inspSimple + inspComplex) / 2;
  const annualSavings = inspectionsAvoided * avgInspectionCost;
  const netYear1 = annualSavings - annualSaaS;
  const upfront = hardwareCost + installation + tagsTotal;
  const monthlyNet = (annualSavings - annualSaaS) / 12;
  const paybackMonths = monthlyNet > 0 ? upfront / monthlyNet : Infinity;
Volume tier: 10–99 points
Sensor
$425*
per monitoring point
SaaS
$12*
per point / month
Tags / sensor
selectedTags.length > 0 ? "$" + tagsCostPerSensor : "—"*
1 × event trigger
selectedTags.length === 0
                        ? "no tags selected"
                        : selectedTags.length === 1
                          ? selectedTags[0].qty + " × " + selectedTags[0].name.toLowerCase().replace("wired ", "")
                          : selectedTags.map(t => t.qty + "× " + t.name.toLowerCase().replace("wired ", "")).join(" + ")
10–99 points
100–249 points
250–499 points
500–999 points
1000+ points
*Pilot-stage indicative pricing. The CSO sensor variant and ATEX-certified tag housing are in active development — final commercials are agreed jointly during launch-partner onboarding.
Pilot Impact — Directional
Hardware Cost
const hardwareCost = points * sensorCost;
50 points · $425/sensor
Installation
const installation = points * installCost;
$600/point
Annual SaaS
const annualSaaS = points * saasPerPointMonth * 12;
$12/point/month
Inspections Avoided
const inspectionsAvoided = totalInspectionsPerYear * (avoidedPct / 100); /yr
of fmtInt(totalInspectionsPerYear) total
Annual Savings
const annualSavings = inspectionsAvoided * avgInspectionCost;
netYear1 >= 0 ? "Year 1 net positive" : "Year 1 net"
const netYear1 = annualSavings - annualSaaS;
Net savings Year 1
Payback Period

Number.isFinite(paybackMonths) && paybackMonths < 240 ? paybackMonths.toFixed(1) months : —

const paybackMonths = monthlyNet > 0 ? upfront / monthlyNet : Infinity; months
At $sensorCost per sensor tagsCostPerSensor > 0 ? " + $" + tagsCostPerSensor + " in tags per point" : "", a single avoided inspection ($avgInspectionCost.toFixed(0) avg) covers a meaningful share of the per-point hardware cost.
Like the directional numbers? Use them as the starting point for a pilot conversation.

Become a launch partner →

These figures are directional, based on the inputs you provide. Pricing reflects the volume tier for your point count automatically. The wired-tag variant and ATEX-rated CSO sensor housing are in active development — launch partners co-shape the configuration during onboarding.
const fmtUSD = (n, digits = 0) =>
  "$" + (Math.round(n * Math.pow(10, digits)) / Math.pow(10, digits))
    .toLocaleString("en-US", { minimumFractionDigits: digits, maximumFractionDigits: digits });
const fmtInt = (n) => Math.round(n).toLocaleString("en-US");

Frequently asked

Honest answers about what's mature and what we're building.

The launch programme is a partnership, so the questions utility teams ask us are partnership questions. Here's how we usually answer them.

What's mature today, and what are we co-developing?
The W-Sensor hardware, the Mimir cloud platform, the edge AI inference pipeline, and the vision-model training methodology are all in production today — deployed at VA SYD and tested in pilot at EMASESA. The CSO-specific deltas — wired wake-on-trigger tags, ATEX Zone 2 housing, sewer-gas-resistant enclosure, and CSO-specific AI models trained on launch-partner site data — are in active development. Launch partners shape the final spec for those deltas.
Why wired tags instead of wireless?
For CSO events, the moment of trigger is the moment that matters. A wireless tag using BLE polling on a low duty cycle could miss the start of an event by minutes. Wired tags wake the W-Sensor instantly when a float or threshold is crossed — image capture is timed to the actual event start, not whatever the next scheduled poll happened to be. For regulator-grade evidence, that timing precision is the difference between a defensible audit trail and a noisy approximation.
What does

Ready to shape the next generation of CSO monitoring?

We're aiming to bring on a small number of launch partners through 2026. If you're scoping a CSO monitoring strategy and want first-mover positioning, let's have a conversation. Quick fit check, no commitment — we usually reply within one business day.