AWS Generative AI-SDLC Companion

November 23, 2025

About the Customer

Customer is a leading U.S.-based insurer providing auto, home, life, and small-business coverage to millions of customers nationwide. Founded in 1928, the company operates through a network of exclusive and independent agents. Customer is known for its strong customer service, broad product portfolio, and emphasis on risk management. ​

The Challenge

The goal of this solution is to leverage the capabilities of Generative AI (GenAI) agents in modernizing enterprise ETL workflows by automating the migration of legacy ETL tools like Informatica PowerCenter, SSIS, and IICS to a modern, scalable, and AI-augmented stack powered by DBT. This initiative is aligned with the client’s digital transformation goals to reduce manual development, increase efficiency, and enable AI-driven governance and observability. ​

The Solution

The architecture consists of event driven services integrating Analyzer, Rally API orchestrators, and DBT code generation logic. The orchestrator agent manages hand-offs, ensuring outputs of Analyzer flow to Rally ticketing, then drive DBT script production.​

Key Objectives

  • Modernize legacy ETL pipelines using LLM-powered agents.​
  • Implement AI-led transformation, mapping, and orchestration in DBT.​
  • Leverage GenAI to detect, generate, and validate ELT code.
  • Establish scalable architecture for future migration programs.​
  • Improve efficiency, performance, and governance of ETL workloads

Core Services

  • Kubernetes Pods for GUI, Batch, Multi-Agent Services​
  • FastAPI (AutoGen 0.4) with Guardrails (PII)​
  • Logging via CloudWatch and Prompt Hub​
Download the Case Study Document

What do you think?

Related articles

Contact us

We Measure the Success by the value delivered to Your End Users

We’re happy to answer any questions and help you determine which of our capabilities best fits your needs. 

Your benefits:
What happens next?
1

We schedule a call at your convenience 

2

We do a discovery and consulting meeting 

3

We prepare a proposal 

Let's Connect
top