Energy & TechCase study

A charging network that runs itself.

Amperly operates an EV charging network on a platform where billing, load management and most support conversations are fully automated.

Network status

1,284 chargers live

⚡⚡⚡⚡⚡⚡⚡

Automated

Billing & reconciliation

ML-driven

Load management

Continuous

Automated operations

Voice AI

Driver support

The challenge

Hardware at national scale, operations stuck at startup scale.

Every new charging site added manual work: commissioning, billing reconciliation, grid-load juggling and a support queue full of “why did my session stop?” calls. The ops team was scaling linearly with the hardware — the economics weren't.

What we built

Automation across the whole session lifecycle.

We built Amperly's operations platform end to end: automated billing and reconciliation, ML-driven load management that balances demand against grid constraints, predictive maintenance on charger telemetry, and Grove voice agents handling the majority of routine support calls across multiple languages.

▸Automated billing & reconciliation

▸ML load management & smart scheduling

▸Predictive maintenance on telemetry

▸Grove voice agents for driver support

▸OCPP fleet integration layer

▸Rootwatch predictive monitoring

The outcome

Reliability up, tickets down, ops team flat.

Session reliability stays consistently high while support tickets dropped meaningfully — with voice agents resolving most routine calls end to end. The network has scaled significantly with no corresponding growth in operations headcount.

At a glance

Industry

Energy · EV infrastructure

Engagement

Accelerator → Dedicated team

Scope

Operations platform + voice AI

Team

Embedded pod

Status

Live · managed under an agreed SLA

GoKubernetesKafkaTimescaleDBGroveRootwatch