Those of us who follow current events will have observed that there is a topic that is gaining speed: how companies are investing in Artificial Intelligence to improve their efficiency and reduce costs in the long term. In fact, reference cases are already beginning to be documented in the Spanish market, such as that of a well-known supermarket chain that has reported savings close to 94% thanks to the development of its own search engine based on Artificial Intelligence.
The data reinforces this perception. According to the latest study carried out by IDC for SAS, 86% of Spanish companies claim to already use generative AI in their daily operations. However, the same report puts a relevant paradox on the table: only 40% declare to have AI truly integrated at the organizational level, that is, operating transversally and providing value in multiple departments. A gap that invites us to reflect on the difference between adopting technology and embedding it in the core of the business.
Another sign of this accelerated, but still poorly integrated, adoption is the Vibe Coding: code generation assistants based on natural language instructions, an accelerator of the development process that helps junior profiles or those with less professional experience reach levels of efficiency that, until recently, were reserved for the most experienced professionals.
However, this immediate accessibility is not without risks and opens up new issues related to security and control.
Resources that are easily available to everyone at a low cost and that speed up our tasks, who wouldn’t take advantage of it? Innocent, well-intentioned behavior that can trigger a critical security issue. We cannot forget that companies handle confidential data, corporate strategies and competitive information that is essential for their success in the market. This information sent to an AI that is in the hands of third parties, and in many cases, without technical supervision or audit, can cause losses greater than the savings achieved. It is a mistake to change corporate control for the ease and immediacy of an external technology, and although we think it is controlled, perhaps it is not so much: according to a report published by MIT in 2025, 90% of workers use LLMs and chatbots on a regular basis, but only 45% of companies declare having purchased LLM subscriptions.
If the objective is for AI to be an engine of real competitiveness and not a trap that ends up giving us more problems than it solves, the industry should define an intelligent strategy, where AI can provide the value it has in a context of trust and security.
To achieve this, companies must understand the importance of AI being an integrated engine in corporate systems, such as, for example, with strategies that do not bring data to AI, but rather bring AI to where companies’ data and their processes reside. When an intelligent assistant operates natively within an analytical and governed corporate platform, the paradigm shifts completely, because the assistant automatically inherits the company’s business rules, security permissions, and historical context.
In this way, effective enterprise AI is much more than vibe coding, it is understanding and processing the company’s massive data, analyzing it, describing it, presenting it in an interpretable way, and developing and executing complex analytical processes. In essence, it democratizes the use of artificial intelligence throughout the analytical process, making it more accessible and encouraging collaboration between diverse profiles.
Effective enterprise AI is much more than vibe coding, it is understanding and processing the company’s massive data
We are already convinced that artificial intelligence can add a lot of value, the next step, the one that converts investment into real productivity, consists of integrating it into internal processes, maintaining business logic, security and established protocols. Because we are talking about something that goes far beyond an assistant that helps write code. Corporate systems are complex and encompass multiple functions throughout the analytical process and having an AI that accompanies that journey, in a governed and native way, will, without a doubt, be what makes the difference.
Cristina Pérez, AI Customer Advisory at SAS
