The intelligence that anticipates and smooths construction transport
EVA uses the real data of your sites and your transport — missions, rotations, field constraints — to help your teams anticipate better. EVA proposes and informs: your teams always keep the decision.
EVA reads your flows before they jam
Fed by the platform’s real data, EVA spots the day’s tensions and proposes a concrete adjustment. Your dispatchers keep the decision.
Planning automation
EVA proposes the order of missions, groupings and slots from the orders. Your dispatchers validate, the unexpected gets reabsorbed.
Logistics still run by phone
The transport of waste, materials and skips by truck still relies largely on manual planning and phone exchanges. These flows are variable, constrained and frequently disrupted: intelligent automation becomes necessary to stabilise and optimise the whole system.
Automating and securing truck flows
The right truck, in the right place
Automatic selection of the type of truck suited to each order (hook-lift, crane, specific capacity) and assignment according to its availability, real location and technical constraints, to reduce unnecessary journeys and improve the execution rate.
A schedule that absorbs the unexpected
Delays, cancellations or new requests are integrated immediately: the system recalculates consistent scenarios without disrupting all the rounds, and maintains operational stability despite the unexpected.
Evidence that checks itself
Vision models analyse the photos taken by drivers to detect skips and identify the waste or materials transported, for more reliable proof of execution and compliant flows.
From request to invoice, a single flow
An architecture links the request, planning, field execution and invoicing in a continuous data flow — fewer manual exchanges, fewer administrative errors, a more reliable logistics process.
What EVA changes, concretely
A multi-agent architecture
To achieve these goals, EVA combines algorithmic optimisation, predictive analysis and automated processing of field data, within a multi-agent architecture capable of orchestrating truck flows in real time.
Assign and recalculate
They process requests and operational constraints — truck type, capacities, time windows, distances, mission sequencing — to propose a consistent assignment, recalculated at every change.
See and verify
They analyse photos and data collected in the field: skip detection, identification of waste or materials, verification of the compliance of evidence, to reduce manual checks.
Orchestrate the whole
It centralises events — new order, cancellation, photo received — identifies the agents concerned and triggers processing, guaranteeing overall consistency and the traceability of decisions.
From prototype to autonomous intelligence
EVA evolves step by step, from research to concrete application. Each year marks a stage towards a smarter, more autonomous system integrated into operations.
Foundations
- Set-up of the architecture and the first intelligent agents
- Centralisation of field data from Nereva
- First prototypes validated on real cases
Field intelligence
- Training of detection agents to identify waste
- Integration of the models into the mobile application
- Optimisation tests with partner logisticians
Cooperation and integration
- Collaboration between agents to enrich data
- New variables taken into account, such as the weather
- Deployment of EVA in a real environment, in an assistant role
Autonomy and explainability
- Agents able to adjust schedules in real time
- Predictive models to anticipate demand
- Explainable and traceable decisions, for more trust
EVA, in the back office and in the pocket
The anticipation panel on the operations side, the assistant on mobile.
Ready to connect your sites and your trucks?
Sign-up is free for 30 days: no commitment, no minimum term.