Novus Hi-Tech enhances Fleet Management System using Amazon Bedrock

India recorded 480,583 road accidents and 172,890 road deaths in 2023, according to the Ministry of Road Transport and Highways. With unsafe driving behaviors continuing to contribute to road deaths in India, Novus Hi-Tech identified an opportunity to help fleet operators act faster and more consistently.

Novus OmniFleet already gave operators visibility into unsafe driving behavior through onboard cameras, GPS, and vehicle sensors. However, as fleet deployments grew, responding to alerts still depended on follow-up from command-center teams. That created operational pressure and made it harder to scale timely, consistent safety responses across fleets.

Novus Hi-Tech set out to close that gap. The company chose Amazon Bedrock to build a system that could assess events in context, determine the next best action, and track whether drivers responded. The company also sought to automate routine follow-up while keeping people involved for validation and higher-risk cases.

Novus Hi-Tech enhanced FleetGPT to manage safety interventions more consistently across expanding fleet deployments. The company automated routine follow-up, applied safety workflows at scale, and kept human teams focused on validation and higher-risk cases. This shifted the operating model from AI-assisted monitoring to AI-driven orchestration and in one of Novus's largest dangerous-goods logistics programs, it contributed to a 55% reduction in drowsiness alerts per 1,000 km and a 46% annual reduction in road-behavior violations.

FleetGPT uses Amazon Bedrock to bring video, telemetry (speed, braking, acceleration, GPS), and driver behavior data together in a single reasoning layer. Each event is evaluated against driver history, operating conditions, and fleet standard operating procedures before the system determines the appropriate response. Within Amazon Bedrock, Novus Hi-Tech uses Anthropic's Claude Sonnet for complex reasoning and Claude Haiku for fast, lower-latency actions. The company also uses Amazon Titan embeddings to help FleetGPT retain context across interactions. Human teams remain the validation layer; FleetGPT handles follow-up, coaching delivery and reinforcement to lock in safer behavior over time.

For real-time data ingestion, Novus Hi-Tech uses AWS IoT Core to connect and manage edge devices securely at scale, and Amazon Kinesis Video Streams to support peer-to-peer communication and video streaming to the cloud. Together, these services ingest signals from advanced driver assistance systems, driver monitoring systems, GPS, fuel sensors, cargo systems, weather inputs and video feeds.

Novus Hi-Tech also uses Amazon Simple Storage Service (Amazon S3) for object storage, Amazon DynamoDB for low-latency serverless data access, and Amazon Aurora for high-performance relational workloads. Together, these services hold the fleet data and historical context that inform FleetGPT's decisions. AWS Lambda runs standard operating procedures automatically, and AWS Step Functions orchestrates the multi-step workflows that span fleet operations.

The company worked with the AWS Prototyping and Cloud Engineering (PACE) team during development. Through this collaboration, Novus Hi-Tech validated its approach and moved from concept to production-ready implementation in under 4 weeks. The company estimates that a comparable internal build cycle would have taken 4 to 6 months.

With Amazon Bedrock now part of FleetGPT, Novus Hi-Tech can extend agentic AI across more workflows, strengthen driver coaching, and support more intelligent decision-making as customer needs evolve. 

The company is also exploring multilingual voice-based agents that can communicate with drivers in their regional languages. This extends FleetGPT in a practical way across India's diverse transport network.

“With Amazon Bedrock and Amazon Bedrock AgentCore powering our agentic AI deployment, FleetGPT now acts on risk – not just flags it. We can extend safe operations across a growing fleet without scaling our command-center proportionally, and that is the shift that changes the economics of fleet safety," said Anuj Kapuria, Managing Director and CEO, Novus Hi-Tech.