AWS ML BlogFriday · May 29, 2026FREE

From data overload to actionable insights: How Verizon Connect scaled agentic AI to 100,000 users

Verizon Connect developed an agentic AI solution on AWS to address data overload from its fleet management platform, serving 100,000 users daily. The system uses AWS services including Amazon SageMaker, AWS Lambda, and Amazon Bedrock to process telematics data and generate actionable insights. Key architectural decisions included a modular microservices design and a multi-agent orchestration framework that delegates tasks to specialized AI agents. Implementation challenges involved ensuring low latency for real-time data processing and maintaining accuracy across diverse fleet operations. Measurable results include a 40% reduction in time spent analyzing data and a 25% improvement in driver safety incidents. The solution is now in production, with Verizon Connect planning to expand its capabilities to additional use cases. The blog post details the technical architecture, including the use of Amazon Kendra for retrieval-augmented generation and Amazon Comprehend for natural language processing.