Monta in Denmark has developed a layer of AI for its charger system software, writes Nick Flaherty.

As EV charging networks expand rapidly, operators face a growing gap between scale and operational capacity. Troubleshooting failures, interpreting fragmented data and maintaining high reliability increasingly rely on scarce technical expertise and manual workflows that do not scale linearly, particularly with fast deployment of charging infrastructure.

Monta AI continuously analyses network operations in the background, surfacing insights, anomalies and recommendations proactively. For example, pushing a firmware update to fix a reliability problem, and to detect and instantly act on fraudulent activity. It also allows operators to ask natural-language questions across operations, performance, pricing, expansion and energy, such as: “Based on EV-equipment in the area, and existing stations from the competition, could you suggest 10 potential sites to review?”

The AI layer is embedded directly into the Monta platform to turn complex operational data into clear, decision-ready insight that can be acted on immediately by any team member, not just specialists. The AI framework is trained on one of the industry’s richest operational datasets, spanning more than 260,000 connected charge points, 3 million monthly charging sessions and 14,000 support requests per month across a unified global platform.

Today, a single failed charging session can require hours of manual investigation across OCPP logs, firmware versions, payment data and hardware documentation – often handled by a small number of experts. As networks grow, this operating model becomes costly, slow and fragile, making it impossible to deliver the reliability needed for this critical infrastructure of the future.

Monta AI addresses this bottleneck by synthesising scattered operational signals into a single intelligence layer to diagnose, explain and recommend actions, while keeping human operators in control of execution.

“The real opportunity with AI isn’t just doing existing tasks faster – it’s enabling operators to do things that were previously impossible. The industry is asking operators to deliver near-perfect reliability with increasingly complex systems,” said Casper Rasmussen, CEO and co-founder of Monta. “Monta AI is built to take pressure off teams. You don’t manage complexity through tools anymore – the system understands what’s happening across the network and surfaces proactively the insight you need to act or initiatives you didn’t even think about”.

An example is that one operator saw a DC charger’s success rate would increase from 31.2% to 98.3% after Monta AI identified a firmware mismatch causing repeated failures.

The AI can identify root causes of failed charging sessions using live OCPP logs, firmware data, payment context and user feedback, and guide teams on priority actions. Network performance optimisation detects recurring issues and performance patterns across chargers, sites and regions to improve overall reliability and uptime. Energy and load intelligence explains smart-charging behaviour, energy constraints and load management factors affecting network performance.

“This isn’t AI in theory,” said Rasmussen. “Monta AI builds on intelligence that already operates at scale. The difference is that we’ve made it accessible to anyone in the organisation, not just engineers.”

In the near term, Monta AI aims to provide operators with AI-powered support to troubleshoot charge points, understand performance issues and manage firmware updates. By analysing operational signals in real time, the AI detects anomalies early with workflows such as maintenance actions or firmware rollouts being partially automated.

Longer term, Monta’s ambition is to support autonomous charging operations, with software orchestrating fault resolution, optimisation and network coordination end-to-end.

 Click here to read the latest issue of E-Mobility Engineering.