From theory to practice: real AI cases

In the second episode of Genius Talks, our host, Alcebíades Araújo, AI Specialist at Squadra, mediated a conversation between Romulo Cioffi, CIOO and Chief AI Officer at Squadra, and Tatiana Oliveira, CEO of AI Brasil, about how to move beyond AI experimentation, with examples of practical applications of artificial intelligence in mission-critical systems and various other business contexts.

Want to watch the full episode? Available on major platforms: YouTube, Spotify, and Apple Podcasts.

AI driving hyper-productivity in mission-critical system modernization

One of Brazil's largest logistics and railway operators had a mission-critical system for optimizing high-traffic rail and sea transport that was no longer keeping up with business demands. The challenge was:

• Over 400,000 lines of C++ code;

• Lack of supporting documentation;

• A system over a decade old with no technical personnel possessing deep knowledge of the available legacy code.

Modernizing the legacy system became imperative to eliminate risks and build a new, solid technological foundation for scalability. Starting with technological modernization, which maintains the system's functional characteristics, the execution was divided into four main stages:

• Discovery: use of LLMs for static analysis and domain clustering to identify complexity points and extract business rules hidden within the 400,000 lines of C++ code;

• Design and modernization pipeline: creation of a new, modern architecture with the system divided into modules for gradual migration;

• Build: Jack assisted the development team in securely converting the C++ legacy code to modern technologies like C# and React;

• Comparative QA: identical scenarios were configured in both the legacy and modernized systems to compare rail schedules, routes, and compositions side-by-side. Any divergence was classified and iteratively corrected until the diagrams converged perfectly;

• Deploy: each module is being implemented in a hybrid operation with the legacy system, allowing for gradual migration as we advance in the overall functional evolution of the system.

"They had an old system, built 15 years ago, that worked well but was a risk to business growth. Through our AI agent network, we extracted business rules from the code, created a new architecture, developed the new system, and deployed it rapidly." - Romulo Cioffi

The legacy modernization process through Genius allowed the operator to build a solid and scalable technological foundation, eliminating risks and preparing the business for continuous functional evolutions starting in 2026. Furthermore, the use of Genius unlocked four fundamental benefits:

• Hyper-productivity and operational efficiency: automation of repetitive tasks and intelligent resource utilization, drastically reducing the time spent by business areas and the time to implement the modernized system;

• Greater accuracy in decisions: improved precision in processes and traceability of AI decisions enhance the final software quality;

• Innovation and competitive advantage: transforming a legacy system into a strategic asset, generating new business capabilities;

• Scalable digital transformation: real technological integration with business processes, enabling rapid adaptation to the market.

Modernizing legacy systems with AI is not just about updating; it's about regaining the capacity to innovate. With Genius, what was once a bottleneck becomes the foundation for sustainable growth, quickly, securely, and scalably.

Claim estimates in minutes with AI

In the automotive and insurance context, Tatiana Oliveira cited a project that was a highlight at CEIA-UFG, the Center of Excellence in AI at the Federal University of Goiás, which used generative AI to read and interpret photos of claims. An analysis that previously took days or weeks to complete is now resolved in minutes, with a near-zero error margin in the generated estimates.

The secret? A journey of almost two decades capturing images and structuring vehicle parts libraries. GenAI was the final accelerator for a robust and proprietary database.

"To reach this level, the journey involved years of capturing images from various claims to perform a comprehensive data aggregation, leading to what the solution is today." - Tatiana Oliveira

The human role: from executors to curators

AI does not replace humans but repositions them. And our new role is that of curators. To this end, certain characteristics are observed:

• Critical thinking: it is the human's responsibility, as a curator, to validate and integrate what the AI produces into the business context;

• Holistic view: the overall view of the context is the new requirement for us to perform intelligent orchestration of business processes through AI;

• Connection between business and academia: the private sector, academia, and government are essential in building a Brazil that moves from being a consumer of foreign technologies to amplifying the use of AI.

This changes roles, it changes the market. The future of AI is as a cognitive amplifier, freeing human talent for activities requiring higher consciousness and creativity, while the machine handles scale and speed.

Do you agree? Want to watch the full episode? Available on major platforms: YouTube, Spotify, and Apple Podcasts.