Indra Unifies EV Charging Data on Databricks

Indra Renewable Technologies, a fast-growing UK smart charger manufacturer, migrated its data operations from a complex web of Azure tools to a single Databricks platform. Previously, the company relied on Cosmos DB, Synapse, Azure Data Lake Storage, Azure Functions, and Power BI. Under the guidance of CTO Matthew Noonan and Data Engineer Meghana Ganatra, Indra adopted a medallion architecture. This setup ingests raw telemetry into a bronze layer, processes it in silver using PySpark, and publishes business-ready gold tables cataloged via Unity Catalog.

The architectural shift allowed Indra to replace three legacy Azure functions with a single fleet pipeline that serves three clients on a daily schedule. This change yielded a 60% to 70% performance improvement over the legacy functions. Overall, the migration from Synapse to Databricks delivered an 80% to 90% reduction in business costs alongside a 90% savings in storage costs. Furthermore, the cost per active user plummeted from £1.30 in January to just 45p in May, while query latency dropped from 24.4 seconds to 3.5 seconds.

To eliminate manual reporting previously done in Excel, Indra deployed automated AI/BI dashboards powered by a serverless SQL warehouse. These tools track real-time metrics, such as a Charger Uptime Analysis dashboard that monitors 14,975 devices with an average uptime of 90.25%. To democratize access, Indra embedded Genie One agents directly into the dashboard canvas. These conversational agents allow non-technical staff to query telemetry, device firmware, and vehicle data using natural language, automatically generating SQL to answer questions in minutes rather than days.

By centralizing its data estate, Indra has established a scalable foundation for future real-time operations. The company plans to expand its use of Delta tables, Lakeflow pipelines, and streaming data to further optimize grid compliance and charger design.