5X FASTER DEVELOPMENT FOR ONE OF BRAZIL’S LARGEST INSURANCE PROVIDERS

Agentic workflows accelerated software development by 5x, enabled code coverage of over 80%, and increased engineering throughput to an average of two stories per day per developer.

INTRODUCTION

Modernizing legacy platforms is one of the most complex challenges organizations face, especially when decades of business knowledge, evolving requirements, and limited documentation are embedded within critical systems.

One of Brazil’s largest insurance providers faced this challenge while evolving a healthcare platform that had supported key business operations for more than 40 years. As demand for new capabilities continued to grow, engineering teams needed a way to accelerate delivery, improve code quality, and reduce the effort required to navigate a highly complex technology landscape.

To support this transformation, CI&T introduced an agentic engineering approach powered by CI&T Flow. By combining intelligent workflows, business context, and engineering best practices, teams streamlined refinement, development, and quality processes—accelerating modernization efforts while preserving the knowledge and stability built over decades of operation.

The result was a faster, more scalable delivery model that sped up development cycles by 5x and established a foundation for future modernization initiatives.

THE CHALLENGE: ACCELERATING CHANGE IN A LEGACY ENVIRONMENT

A highly complex platform with decades of embedded business logic made it difficult to reduce delivery timelines, improve predictability, and scale modernization initiatives.

Modernizing a Highly Complex Legacy System

The platform had been in operation for more than four decades and contained extensive business logic accumulated over years of evolution.

Limited documentation and deep technical complexity made it difficult for teams to quickly understand dependencies and safely implement changes.

Knowledge Gaps During Refinement

Teams frequently encountered challenges translating business requirements into development-ready work.

Missing information, undocumented rules, and refinement gaps often resulted in rework, delays, and unplanned development effort.

Long Development Cycles

Complex dependencies and manual engineering activities increased delivery timelines and reduced the team's ability to respond quickly to business demands.

THE SOLUTION

CI&T introduced an AI-powered engineering framework designed to support the entire delivery lifecycle—from discovery and refinement through development and testing. Powered by Flow, the solution combined agentic workflows, development accelerators, and business-context enrichment to improve both delivery speed and quality.

AI-Driven Discovery and Refinement

The initiative leveraged AI agents to support both business and technical refinement activities.

Project repositories, discovery artifacts, and design assets were integrated into the workflow, allowing agents to analyze requirements, identify dependencies, and generate more complete implementation plans.

This helped reduce ambiguity and improve alignment between business and engineering teams.

Accelerated Development with AI Agents

Specialized development agents supported coding, code review, bug fixing, linting, and unit test creation.

Rather than replacing developers, the agents handled repetitive implementation tasks, allowing engineers to focus on validation, architecture decisions, and complex business logic.

Quality Built Into the Process

To ensure consistency and maintainability, guardrails and engineering standards were embedded directly into agent workflows.

This helped improve code quality, increase test coverage, and reduce the risk associated with changes to a critical legacy platform.

THE IMPACT

The adoption of agentic engineering practices delivered measurable improvements in both productivity and software quality.

DEVELOPMENT ACCELERATION

• 5x faster development and code review cycles

• Average delivery rate of two stories per developer per day

• Faster execution of prioritized work within the current planning cycle

IMPROVED QUALITY

• More than 80% code coverage achieved through AI-assisted development practices

• Greater focus on validation, edge-case testing, and business scenario verification

• Reduced effort spent on repetitive coding activities

FASTER REFINEMENT

The organization also began applying AI-powered upstream agents to technical and business refinement activities.

• Expected refinement cycles completed in less than one week

• Approximately 3x faster refinement process compared to previous delivery cycles

TECHNOLOGIES USED

POWERED
BY FLOW

CI&T Flow Enterprise AI Management System that integrates intelligent agents into software delivery workflows, helping organizations accelerate development, improve quality, and increase engineering productivity.

AGENTIC ENGINEERING FRAMEWORK

A coordinated set of AI agents supporting discovery, refinement, development, testing, code review, and issue resolution activities throughout the software lifecycle.

AI-DRIVEN AGENTS

Specialized agents designed to automate repetitive engineering tasks such as code generation, review, bug fixing, linting, and unit test creation, enabling teams to focus on higher-value activities.

CONTEXT-AWARE REFINEMENT

Integration of business requirements, design assets, technical documentation, and project repositories to provide AI agents with the context needed to support more accurate refinement and planning activities.