Anti-fraud reliability: how Ecomilhas guarantees every trip is real

When you engage employees to decarbonize daily commutes and business travel, trust is everything: every reward on the card and every tonne of CO₂e in the inventory is only worth something if the trip behind it is real. That is why Ecomilhas never relies on a single signal — every trip goes through a combination of independent criteria, assessed in real time by Emília, the Ecomilhas AI, which together form a reliability index. This evidence engineering is what makes us the benchmark in reliability for decarbonization through engagement.

Meet Emília, the Ecomilhas AI

Emília is the intelligence that analyses every trip. She is a neural network with a mature convolutional layer, refined through machine learning on more than 20 million trips submitted to the app. It is this accumulated experience — millions of real examples of cycling, bus, metro, train, electric car, walking and carpooling — that gives Emília the repertoire to recognize, in seconds, what is an authentic trip and what is an attempt to game the system.

The reliability index: a consensus of evidence, not a single test

The reliability of a trip does not come from one isolated test — it is a score. For every trip, Emília combines several independent criteria, each looking at the journey from a different angle. When they all point to the same story, the trip is approved. When one signal is out of step with the rest, the score drops and the trip is reviewed or discarded. It works like a panel of experts inside a single AI: one angle can be fooled, the group cannot.

The layers Emília cross-checks on every trip

• Physical profile of the trip — speed, acceleration, route shape, adherence to roads, bike lanes or rail tracks, stops and restarts.
• Origin and destination — a real trip takes a person from one point to another.
• GPS signal integrity — authentic signals have a signature of their own, hard to forge.
• Connection data — continuity of the trace and network behaviour along the entire route.
• Identity — the right person being rewarded for their own action.
• Historical pattern — how that trip compares to the user's typical behaviour and to similar trips already validated.

From this combined picture, Emília classifies the most likely mode of transport and measures how coherent the trip is. A “cycling” trip at 90 km/h on a highway, for example, does not pass.

How Emília catches each type of fraud

Treadmill movement: origin and destination do not add up

Someone running on a gym treadmill generates plenty of movement — but never leaves the spot. By cross-checking origin and destination, Emília sees that there was no real displacement between two points: a lot of activity, no trip. With no genuine displacement, there is no reward and no accounting.

Fake GPS: duplicate signals give it away

Fake location apps try to “teleport” the device to simulate a clean route. But forged signals leave traces: Emília detects duplication and inconsistency in the GPS signal — when the same signal source shows up in ways the real world would not produce. That duplication signature exposes the simulation, and the trip is blocked.

Identity fraud: FaceID verification

A perfect trip counts for nothing if the wrong person is being rewarded. FaceID verification ties the trip to the employee's biometrics, ensuring that whoever accrues balance is in fact the person who travelled — and closing the door on borrowed or shared accounts.

Disagree with a rejection? There is a second review

No system is infallible, and Ecomilhas handles that with transparency. If you believe a trip was rejected unfairly, you can request a second review, deeper and more thorough, directly in the app — Emília re-examines that trip at a greater level of detail. The outcome of this second review is final: there is no appeal against reviews. This re-analysis is an exclusive benefit for Ecomilhas Green Card subscribers.

What happens when something does not add up

Trips that fail any critical check are discarded: they do not become balance on the card and they do not enter the auditable inventory. Only what is consistent across every layer is rewarded and counted — like a silent gatekeeper separating the real from the forced, with no friction for those acting in good faith.

Why companies and employees love this

• Fair rewards — those who genuinely travel clean are recognized; those who try to cheat are not.
• A reliable inventory — avoided CO₂e data is clean from the start, trip by trip, ready for the GHG Protocol, CDP and CSRD with an auditable trail for each journey.
• Engagement that lasts — clear, fair rules sustain participation over time, without that feeling that “the system is easy to trick”.

Privacy first (LGPD)

Emília assesses technical signals from the trip and the identity verification, always in compliance with the LGPD (Brazil's data protection law): the company accesses aggregated emissions data — never an individual's real-time location.

It is this combination of criteria — read by Emília, the mature Ecomilhas AI, trained on more than 20 million trips — that makes Ecomilhas the benchmark in reliability for engaging employees in the decarbonization of daily commutes and business travel. Here, every low-carbon credit is, in fact, low carbon.