AI-Charge

Pilot · Vietnam

An operating layer for EV charging networks

Charge points send telemetry. AI-Charge stores every event, keeps a live picture of each connector, forecasts when a busy station frees up, flags the ones about to fail, and shows the operator what the week actually looked like.

How the data moves

One path, no manual steps between the charger and the screen.

  1. 01

    Charge point

    Status changes, sessions and faults arrive over a webhook, authenticated per operator.

  2. 02

    Raw log

    Every event is stored verbatim first. Projections can be rebuilt from it at any time.

  3. 03

    Live state

    Stations, connectors, sessions and faults are projected as the events land.

  4. 04

    AI layer

    Freeing-time forecast, failure risk, utilisation and demand-based price hints.

  5. 05

    Cabinets

    Operator dashboard and driver map, updated in real time over a live stream.

What the platform does

Live station state

Per-connector status derived from the operator’s own feed, not from a schedule or a manual update.

Freeing-time forecast

How long a busy station is likely to stay busy, computed from its own session history.

Failure risk

Stations ranked by how likely they are to fail, with the fault history behind the number.

Maintenance alerts

Overheating, current leaks, repeated faults and silent chargers surface as a ranked feed.

Operator analytics

Sessions, energy, downtime and availability over any period, against the period before it.

Price hints

A demand-based tariff suggestion per station. The operator decides whether to apply it.

Two cabinets

Operator

Fleet dashboard, per-station history, analytics over time, maintenance feed and tariff hints — limited to that operator’s own stations.

Operator dashboard

Driver

Map and list of stations, freeing-time forecast, a hold on a charger and a corridor of stations along a route. Runs on the simulated network in the demo.

Driver cabinet

Every operator sees only their own network

Stations belong to the operator that sends them. Reads are limited to that operator on the server, telemetry is authenticated per operator, and the live stream never carries another network’s station to a subscriber.

The demo runs on a simulated network

It is a working copy of the driver cabinet fed by a generator, not by a real operator. No customer data is inside it.

Open the demo

Connecting a network, or looking at the platform?

Write to us — we will show the operator cabinet on live data and talk about what your feed needs to carry.

Write to us · hello@ai-charge.dev