Mercadona has developed and launched Atenea, its own artificial intelligence platform with the aim of helping its technological teams to generate software more quickly and with greater quality. It is estimated that more than 500 engineers are already using this agentic AI platform created to transform and accelerate large-scale software development. With this tool, which not only responds or generates content alongside the user, but can also execute tasks autonomously, the company seeks to gain agility and efficiency in the creation of applications and, thereby, promote the digitalization of its physical processes.
Coordination of specialized agents
Atenea is a solution from Mercadona IT, the department responsible for the company’s digitalization. The platform integrates Mercadona’s knowledge, processes, tools and quality standards, and coordinates specialized agents capable of executing different missions within the development cycle. With this solution, the company seeks to optimize the development of the software that supports the business.
“This proprietary artificial intelligence platform helps technology teams create software faster, with higher quality and, most importantly, integrating the knowledge and context of our company,” explains Sergio Pajares, CTO of Mercadona.
Teams from different areas of Mercadona IT participate in the development and implementation of Atenea, including processes, technology, operations, infrastructure, design or security, among others. In addition, the company has formed a team of “early adopter” engineers, in charge of promoting its implementation in the different development verticals for the business, as well as training and accompanying the teams in the adoption of this new way of working.
According to Sergio Pajares, CTO of Mercadona, “with the development and launch of Atenea, the company seeks to create a new way of developing software, faster, more consistent and aligned with the needs of the business. In this sense, the main challenge has been for the AI to understand how Mercadona works and for Atenea to function as a team made up of different specialized agents, each one prepared to perform a specific task or work with a specific technology.”
The next phase involves extending the use of agents to the entire development life cycle, making the platform more efficient and evolving it so that it can work with AI models from different manufacturers.
An AI that understands how Mercadona works
The differentiating factor is not only in having increasingly capable AI models, but in ensuring that they understand how the company works and act within its technical criteria. “Artificial intelligence is increasingly capable and efficient. But, if clear limits and objectives are not set, it can make mistakes or work in the wrong direction. The important thing is not only to have a very powerful AI, but to ensure that it is aligned with the needs of the company,” says Pajares.
Before tackling a task, Atenea provides agents with information about how Mercadona develops software: its processes, tools, standards, designs and quality criteria. In this way, AI can generate solutions that fit with the company’s real way of working. “AI does not replace technical responsibility; it amplifies it. Professionals maintain control and judgment over decisions, while the platform multiplies their execution capacity,” adds Pajares.
AI that impacts brand efficiency
The company has spent years developing a technological structure that takes advantage of AI when it truly adds value to directly impact the efficiency of its business. Hence the importance, especially in a company of its size, with hundreds of applications and processes, of correctly identifying where artificial intelligence provides a real impact.
The company has been developing a technological structure for years that takes advantage of AI when it truly adds value
In addition to Atenea, Mercadona has other solutions that incorporate AI in different processes. Some of them include the planning and management of vacations and schedules, the automatic recognition of supplier invoices or an own project to manage all product master data with AI.
