Deploying an AVM across 30 lines: real timelines and key factors

"How long does it take?" That is usually the first question asked in an AVM project. The honest answer: between 6 and 16 weeks depending on context. And the variations are almost always explained by the same factors.

The scope that determines everything

For a 30-line network, deployment time depends less on size than on inherited complexity. The determining variables are the quality of existing GTFS data, the diversity of the rolling stock, the number of depots, and the integrations required with other systems such as ticketing, passenger information displays, and scheduling software.

A network with clean GTFS data, a homogeneous fleet, and an available project team can be live in 6 to 8 weeks. A network with fragmented data, heterogeneous equipment, and multiple stakeholders to coordinate can reach 14 to 16 weeks.

The 4 phases of a SaaS AVM deployment

Phase 1: Audit and scoping (1 to 2 weeks). Fleet inventory, onboard equipment compatibility check, mapping of existing data flows.
Phase 2: Network data import (2 to 4 weeks). GTFS import, geographic consistency check of stops, line and service configuration. On a SaaS AVM like Pysae, this phase is handled by the deployment team without any intervention on the client's servers.
Phase 3: Configuration and training (2 to 4 weeks). Driver interface configuration, alerts, punctuality rules. Operations supervisor training (2 to 4 hours is generally sufficient) and driver onboarding (under one day on Pysae Driver).
Phase 4: Staged launch and stabilisation (2 to 4 weeks). Line-by-line or depot-by-depot rollout. GPS coverage rises quickly: on the TeMo Moselle Nord network, 73% of trips were tracked from day one.

What accelerates, what slows down

Acceleration factors: up-to-date GTFS, a designated project team, a progressive rollout, choosing a SaaS AVM with no server infrastructure to install.
Slowdown factors: fragmented data, multiple integrations in parallel, long internal validation processes, a heterogeneous vehicle fleet.

What changes after deployment

The first visible effect is the disappearance of cascading calls to the depot to locate buses. On networks like Keolis Porte de l'Isère, monthly report production time dropped from several days to 15 minutes.

Does your network have 30 lines or more? Talk to our team.