How SSE Airtricity Drives 90% Customer Engagement with AI Energy Insights

SSE Airtricity serves around 750,000 customers across Ireland with electricity and gas, and has made a clear commitment to leading the next wave of energy innovation. With smart meters already deployed across a large share of its customer base, the company had rich consumption data at its disposal. The challenge was turning that raw data into personalized guidance that could shift customer behavior and support its net zero ambitions.

Existing digital tools fell well short. Basic usage visualizations and generic advice—“switch to LED bulbs”—failed to meet the expectations customers had developed from retail and media platforms. SSE Airtricity had explored black-box third-party solutions, but found they offered no visibility into which features customers actually valued or used. The team needed something they could control, iterate on quickly, and measure precisely. For a data and analytics team entering generative AI for the first time, building in-house felt both necessary and daunting.

SSE Airtricity’s data and analytics team built the Enhanced Smart Insights (ESI) Energy Advisor from scratch on Databricks. The platform’s Unity Catalog provided centralized governance for both data and AI models, while Delta Lake underpinned the data architecture. Lakeflow Spark Declarative Pipelines ingested billions of rows of smart meter data—48 daily readings per meter—processing it every night to feed the recommendation models. Claude Sonnet 4.5, accessed through Databricks AI Model Serving, serves as the reasoning engine—generating personalized, actionable advice per customer without requiring infrastructure management. MLflow supplied end-to-end observability through traces and AI judges that automatically reject and regenerate any insight failing quality thresholds for length, tone, language, and accuracy. Custom monitoring apps built with Databricks Apps track cost, model performance, and judge failure rates in real time.

The ESI Energy Advisor soft-launched in mid-December 2025 and exceeded expectations immediately. Within the first month, unique visitors to the insights section rose 12%. Ninety percent of logged-in customers navigate to the insights page and engage with its recommendations; 10% click through to their full personalized insight set. The team went from proof of concept to production in roughly four months—a pace enabled by Databricks accelerators and the evaluation-first methodology that gave business stakeholders confidence before any code reached production.

SSE Airtricity now has a governed generative AI capability in production and is extending it rapidly. Upcoming features include tariff recommendations, a bill analyzer, peer comparisons, and energy disaggregation. The system’s ability to update insights automatically every two months—when bills arrive—gives SSE Airtricity a concrete competitive edge. “A competitor requires customers to wait 120 days after signup to receive insights,” noted Maria Leacy, Analytics Manager. “With us, the day after you get a bill, you can have your insights. It’s a real differentiator.” The platform also positions the company to respond quickly to dynamic pricing, new tariff structures, and emerging flexibility programs that will define the next phase of the energy transition.