Smart Solar Alert System
Mar, 2025
Project background
In 2024, Amergy Solar—one of the U.S.’s leading solar providers with more than 9,000 residential and 300 commercial systems installed—partnered with Kanta Tech to redesign its outdated solar station monitoring and alert system.
The existing monitoring approach treated all power losses identically. A solar station’s natural sunset shutdown triggered the same alert as a genuine equipment failure. As a result, real failures occurring at night often went undetected until daylight, keeping stations offline for hours or even days. Most providers had accepted this inefficiency as inevitable—until this project began
The Challenge
Solar downtime translates directly into revenue loss. Every missed hour of sunlight meant wasted energy and financial penalties. Yet the industry’s core blind spot lay in the inability to distinguish between natural sunsets and genuine station failures.
This was not simply a matter of engineering speed—it was a systemic problem. Managers focused on dashboards, while 98% of actual repairs were carried out by frontline technicians who had never seen the alerts. The system was designed for the wrong audience and could not separate false alarms from real failures.
Our Solution
Kanta Tech reimagined the system from first principles:
Field Research: Gou spent days in the field, interviewing technicians and studying operational records spanning 15 years.
Breakthrough Discovery: The key lay in voltage patterns. Failures produced a sharp drop, while sunsets followed a smooth, gradual decline.
AI Algorithm Design: Using this insight, the team built the first AI-powered model capable of separating sunsets from true equipment failures.
Human-Centered Redesign: Instead of centralizing control with managers, the new system empowered frontline technicians with direct, actionable alerts.
To illustrate the distinction, Gou used a simple but powerful metaphor—tipping a coffee cup to mimic a sudden failure splash, and then showing a slow drip to represent sunset. This vivid demonstration aligned stakeholders across technical and managerial levels
Execution
The system was piloted across three solar sites in New Jersey. Gou led her team to not only deploy the AI model but also redesign adoption practices:
Built tailored demos for managers and technicians to ensure buy-in from both sides.
Introduced a feedback loop that allowed technicians to tag alerts as “false alarm” or “real failure,” enabling continuous system improvement.
Conducted A/B testing to refine both technology and user experience.
Results & Impact
Downtime reduction: Repair response time dropped by over 98%, from nearly a week to three hours.
Financial savings: Prevented a $150,000 penalty in the first three weeks and has since avoided $1M+ in losses.
Scalability: Adopted across all New Jersey sites, with expansion to New York and California underway.
Industry benchmark: Other companies have approached Amergy Solar to license the algorithm—a first in its history.
Cultural shift: The system turned technicians into collaborators and advocates, ensuring sustainable adoption beyond the pilot region.
Looking Ahead
This project established a new industry standard for alert systems in solar energy. More importantly, it demonstrated that successful digital transformation is not only about algorithms, but also about aligning with human workflows. By combining AI precision with frontline expertise, Amergy Solar and Kanta Tech created a model for the future of renewable energy operations.