Case Studies — SERRAO Analytics Deployments

Steel & Metals · Safety Monitoring

Real-time PPE compliance at a steel factory

Client

Confidential

Problem

Manual PPE checks could not cover the full floor across shifts, and violations went unrecorded.

Solution

Computer vision on existing CCTV detected helmet and jacket violations in real time, emailing supervisors a snapshot per event.

~98%Helmet-detection accuracy

~87%Violations caught in pilot

Discuss a similar deployment

Steel & Metals · Safety Monitoring

Loading & unloading safety at a steel plant

Client

Confidential

Problem

Material handling created unsafe proximity to trucks and inconsistent use of safety gear.

Solution

The system detects proximity to trucks during active operations and flags missing safety gear during loading and unloading.

ProximityAlerts near vehicles

GearChecks during handling

Discuss a similar deployment

Distribution Utility · Power & Electricity

Energy-theft detection for a state DISCOM

Client

Confidential

Problem

High AT&C losses with limited ability to target field inspections effectively.

Solution

An anomaly-detection layer over consumption and metering data prioritised probable-theft cases for field verification.

AT&CLoss-reduction focus

TargetedInspection lists

Discuss a similar deployment

Metropolitan DISCOM · Power & Electricity

Tariff intelligence & revenue protection

Client

Confidential

Problem

Revenue leakage from tariff-category misclassification and unverified preferential tariffs.

Solution

AMI-data models detect misclassification and verify preferential tariffs using load signatures (NILM).

AMIData-driven detection

NILMLoad-signature checks

Discuss a similar deployment

Fraud detection and risk management for a state power distribution utility

Public Sector · Power & Electricity

Fraud detection & risk management for a state DISCOM

Client

Confidential · Tamil Nadu power distribution utility

Problem

Revenue leakage across LT service connections, with billing and consumption data too large in scale for manual review — spanning roughly 3.3 crore LT consumers from 2021 onwards.

Solution

Data models built to analyse LT billing and consumption data at scale, generating exception (red-flag) reports that surface probable revenue leakage for targeted follow-up.

3.3crLT consumers analysed

Red-flagException reporting

Discuss a similar deployment

Deployed across safety, power and custom builds

Real-time PPE compliance at a steel factory

Loading & unloading safety at a steel plant

Energy-theft detection for a state DISCOM

Tariff intelligence & revenue protection

Fraud detection & risk management for a state DISCOM

Discuss an AI-ML deployment