← Cobra Propreté · Use case
City of Metz
From perception to management: how the city of Metz turned dirtiness levels into objective data and gained control over its street-cleaning costs
Metz is moving from a subjective approach to data-driven cleanliness.

KEY FIGURES
What Cobra Propreté changed in Metz
25 %
of time saved on audits
Thanks to our frugal AI
200
hotspots objectively mapped
30 of which gone between 2024 and 2025.
150
operational agents
Adoption in less than a day, even by agents most reluctant to go digital.
CONTEXT
Cleanliness under regulatory and budgetary pressure
The core mission of Metz’s urban cleanliness service rests on a trio of elected officials, field agents and residents: implementing the strategy set by elected officials while fostering agent engagement to satisfy residents. Like any local authority with responsibility for cleanliness, the city of Metz faces a twofold challenge:
- Evaluating service quality: IOP audits (Indicateurs Objectif Propreté, a methodology proposed by the French AVPU non-profit) and the management of REP (Extended Producer Responsibility) schemes, including the PLDA (litter-fighting plan) with CITEO and cigarette-butt littering with ALCOME, require regular, usable measurements.
- Keeping resources under control: team leaders and technicians handle unexpected events on top of street furniture management and their daily operational work, in a financially constrained context.
The limits before Cobra Cleanliness
Before the rollout, cleanliness measurement relied on:
- Manual counting using a digital counter.
- Quarterly results: impossible to react quickly to a deterioration on the ground, with waits of up to 3 months to know where to adapt resources.
- Subjective field knowledge: priorities depended on team leaders’ impressions, with no objective mapping of areas.
The challenge
The city of Metz needed to improve how it evaluated the effectiveness of its street-cleaning operations, in order to prioritise them systematically.
THE SOLUTION
Turning the field into objective data to steer action
A rollout that is “child’s play”
The City of Metz deployed Cobra Cleanliness without a heavy IT project:
- No dedicated training: the app spread on the ground, from agent to agent.
- Up to speed in less than a day: just the time needed to test the interface and agree on audit routes.
- Adopted by everyone, the app is described as “child’s play”.
A response to the fear of being monitored
As is often the case, the first resistance was the fear of a staff-surveillance tool. The City answered with transparency:
- Demos and hands-on trials;
- Audits are carried out BEFORE the agents clean up, not after;
- The tool is not intended to replace jobs, but rather to speed up agents’ work.
Everyday use
On the ground, the application adapts to the reality of each space: large areas undergo automated visual analysis, while sensitive or smaller areas, such as the surroundings of schools, are covered by targeted manual analysis.
Beyond the snapshot, Metz relies on historical trend curves and the volume of audits carried out to understand patterns and make trade-offs more easily.
Cobra Cleanliness is also a tool for measuring the effectiveness of investments. Every change of street furniture is evaluated through the evolution of cleanliness in the area concerned. The city of Metz can now prove that its furniture is positioned in the right place, and move it when the data shows it would be more useful elsewhere.
Support from across the local authority
The elected official in charge of urban cleanliness
[Cobra Cleanliness] embodies the city we want: clean and looking to the future.
As deputy mayor for urban cleanliness and waste collection, I fully support its ambitions and encourage the teams to evolve their practices, in a spirit of continuous improvement.
Martine Nicolas, Deputy mayor for urban cleanliness
The IT department
Drawn to the artificial intelligence dimension of the application, the IT department took an interest in it and was able to support the project’s rollout.
RESULTS
Fine-grained data, faster decisions, more efficient audits
From a quarterly zone-level IMS to a street-level IMS
Before the rollout of Cobra Cleanliness, the City of Metz only had a global IMS (Municipal Cleanliness Indicator), calculated by zone and reported every quarter. This rhythm meant waiting up to three months to measure the effect of an action or detect a decline in service quality.
Today, the index is available at street level, continuously: teams can audit a street at any time and react immediately. Measurement is no longer a burdensome reporting exercise, but an operational lever serving everyday decision-making.
“Last week, I carried out 3 audits on 3 streets that are usually very clean. This made it possible to react immediately to this field issue.”
Jonathan Maurice, Studies officer for the improvement of operating resources at the city of Metz
Audits 25% faster
Thanks to automated visual analysis powered by our frugal AI, agents save 25% of their time during audits. This freed-up time can be devoted to more audits or reinvested directly in action on the ground.
Objectively mapped hotspots (PLDA)
As part of the PLDA and cigarette-butt schemes, the city identified 200 hotspots. The proposed dirtiness scale has 4 levels (very dirty / dirty / moderately clean / clean), but its weakness is that it is subjective.
The counts carried out with Cobra Cleanliness make it possible to characterize and calibrate this scale. This allows areas to be compared objectively and to move beyond the inevitably variable impressions of team leaders.
30 hotspots were cleaned up between 2024 and 2025, and the average dirtiness level decreases year after year, proof that corrective actions are paying off.
A model to quantify the impact of hotspots
To evaluate the ROI achieved on hotspots, the city of Metz built a rigorous model.
Each hotspot is first classified into 4 typologies (scattered abandoned waste, waste contrary to collection rules, wild waste mostly consisting of household packaging, and multiple waste items) and positioned on the 4 dirtiness levels. Each level corresponds to a standardized cleaning time as well as an intervention frequency (daily, weekly or multi-year depending on the location). The financial calculation then crosses the number of agents deployed, the number of passes required and the cost of the mechanical means used. This model gives decision-makers a clear vision: every hotspot cleaned up translates directly into working hours and euros saved.
In 2026, the Cobra Propreté approach will improve and consolidate this objective methodology.
Another effect: a meaningful approach and a sense of purpose
Beyond the numbers, the tool produced another effect: a stronger bond between managers and field agents. Regular audits create natural opportunities for exchange and direct dialogue with the teams, on the ground, around concrete situations. This proximity made it possible to re-engage staff and reposition street furniture in the right places, combining agents’ field knowledge with the tool’s data.
The teams realize that data does not replace them: without them, management can do nothing. Cobra Cleanliness becomes a lever for recognition of agents’ work and for collaboration.
IN BRIEF
What Metz gained
| Before | After Cobra Cleanliness |
|---|---|
| Constraining IOP audits, manual counting | Fast, intuitive and versatile visual analysis |
| Quarterly results | Immediate responsiveness (audits on demand) |
| Subjective field knowledge | Objective evaluation of the 200 hotspots |
| Unmeasured effectiveness of actions | Every action (furniture, corrective work, awareness) is measured |
| Management without relevant indicators | Data-driven management, audits 25% faster |
Switch to data-driven cleanliness too.
Cobra Cleanliness lets you objectively assess dirtiness levels, map your hotspots and speed up your audits, with no heavy training or costly IT project.