The Unicaja entity is advancing in the evolution of its software engineering model with the industrialization of code production through capabilities augmented by generative artificial intelligence (IAG). This approach seeks to scale systems development in a context of increasing technological complexity and regulatory demand, preserving internal control over the architecture, security and governance of the technology.
The model is based on the standardization and factorization of development processes, on the use of IAG as a support tool, always with human supervision, and on collaboration with a stable ecosystem of technological partners organized by domains, formed by IBM, NTT Data, Babel, GFT and Scalian, which complement the bank’s internal capabilities and operate in accordance with the criteria and frameworks defined by Unicaja.
Control over technological architecture
This evolution of Unicaja is part of the work that the entity has been developing since the beginning of the 2025-2027 Strategic Plan, aimed at providing software engineering with greater capacity to accelerate the execution of projects, improve efficiency and facilitate the evolution of platforms, fundamentally the most traditional ones, while retaining control over the technological architecture and the governance of the use of the IAG within the bank.
The incorporation of the IAG into the work environments of the Unicaja Technology team is being carried out progressively. In the first phase, Copilot has been deployed as a support tool for certain stages of project execution, such as gathering business requirements, project management and improving the quality and consistency of deliverables. This deployment has been accompanied by a structured change management process, aimed at facilitating the adoption of these new ways of working and generating good practices with an impact on deadlines and productivity.
The central axis of the model is centralized knowledge management, articulated through a proprietary platform developed with the support of IAG models selected and validated by the bank’s AI team. These models have demonstrated high efficiency in understanding and processing both current and legacy programming languages.
Reduce operational and obsolescence risk
The entity’s internal platform, called Rosetta, acts as a common base for new projects and applications in use, providing homogeneous work criteria and orderly and governed access to functional and technical information for internal teams and for the engineering centers of selected collaborators. In this way, it contributes to mitigating operational and obsolescence risk. From this basis, the incorporation of programming tools is considered an immediate step, always with human supervision integrated into the process.
In 2025, the Rosetta embryo was developed through refactoring tests of traditional elements of the Unicaja mainframe platform (JCLs). In these pilots, programming time savings of more than 80% and a level of precision of the refactored code close to 100% were recorded. During 2026, it is planned to incorporate Rosetta, in different waves, applications from different domains together with selected partners, starting with the most legacy technologies and moving towards digital and data platforms. In the current year, approximately 600 business applications currently in production at Unicaja will be incorporated.
The incorporation of programming tools is considered an immediate step, always with human supervision integrated into the process.
“This approach allows us to advance with greater agility in the development of our systems without giving up essential principles such as quality, control and security, in line with the responsible adoption framework of AI that Unicaja has been deploying, with human supervision, integrated governance and regulatory compliance,” said Estrella Botas, general director of Technology and Operations at Unicaja. “Industrialization supported by generative artificial intelligence and an aligned ecosystem of partners, reinforces our internal capacity and prepares us to evolve and modernize our systems with greater efficiency and solidity,” he added.
