AI has been dominating headlines for two years and most companies are already using generative models to create text, images or code. But there are other types of artificial intelligence, such as Agentic AI, which can transform business efficiency and customer service.
Two key reports such as MIT’s “State of AI in Business 2025” and IDC’s “Data and AI Impact Report” commissioned by SAS predict an imminent adoption landscape, with profound implications for Spanish companies. The fundamental question now is: what exactly is this ‘Agentic AI’ that promises so much and what do organizations need to know in order not to be left behind in this technological race?
The Agentic AI Imperative
The report from SAS, a leader in data and AI, reflects that there is an adoption of Agentic AI of 52% globally, and 56.6% in Spain. It also shows that companies are almost equally interested in industries that use “traditional” machine learning AI, and industries that prefer to use Agentic AI and generative AI.
In addition, it is found that 78% of business leaders consider advanced AI as a critical factor in which they fully trust to maintain competitive advantage in the next three years
According to the MIT report, what executives expect from Gen AI is flexibility, a deep understanding of companies’ workflows, minimal disruption and the possibility of improvement over time. This suggests that European companies are aware of the transformative potential and the need to get on this technological train.
Agentic AI has the advantage in this sense, since it is overcoming these challenges: it does not require context every time information is included, it has persistent memory, it learns from interactions and it can manage tremendously complex workflows.
And, while generative AI focuses on creating content, Agentic AI operates as an intelligent companion capable of reasoning, remembering relevant information, making decisions and executing multi-stage actions with minimal human guidance.
As detailed in the practical guide on Agentic AI from SAS, these systems are characterized by their perceptual capacity to understand the environment and available information, followed by deep cognition to reason, plan and process complex information. This understanding allows you to choose the most appropriate action and execute it. Furthermore, continuous learning allows performance to be improved through experience. Agentic AI seeks to be highly specialized and have autonomy in specific domains.
An autonomous and responsible AI
The adoption of Agentic AI is a strategic and organizational issue. From SAS, Amaya Cerezo, Artificial Intelligence & Analytics expert for SAS Spain, ensures that to implement and use Agentic AI productively it is essential to address several pillars. “First, governance is key to ensuring that all agents within the organization are controlled, transparent and aligned with corporate strategy and values. The human must also be kept involved, as human oversight ensures that critical decisions always maintain human judgment when necessary. Finally, it is vital to encourage intelligent decisions, integrating agents seamlessly into routine business operations to optimize processes and improve overall efficiency.”
The Foundations of a Productive Agent: Key Accelerators
The most effective agents are built with specific tools and components that ensure they function optimally. According to Amaya Cerezo, at SAS they have identified and developed key “accelerators” to enhance Agentic AI. Amaya explains: “RAGs (Retrieval Augmented Generation) stand out, which are essential for agents to access and reference specific business data sources, and offer contextual, precise and high-value responses for businesses. In addition, the management and improvement of prompts is crucial to calibrate and refine the instructions that guide agents, improving the quality and relevance of their actions and results.”
Another key aspect is the orchestration of agents, which allows the creation of specialized ecosystems that work in a coordinated manner, where each one plays a specific role in multi-stage tasks. This allows us to maximize the efficiency and complexity of the solutions they are able to offer.
Persistent Challenges and the Need for Preparation
Despite its promise, Agentic AI also faces familiar challenges. 46% of companies at European level find themselves in a dilemma regarding trust in AI, because there is a clear difference between real trust and perceived trust towards this technology.
In this sense, it is concluded that the need for new regulatory frameworks is imperative to guarantee responsible implementation and protect both companies and citizens.
Regarding the top 3 priorities for managing AI projects, European companies highlight that they want to create an AI architecture (51.2%), that learning and reskilling is imperative to manage this technology (44.8%) and create data science and AI teams (41.9%))
Agentic AI operates as an intelligent companion capable of reasoning, remembering relevant information, making decisions and executing multi-stage actions
For its part, the MIT report concludes that ‘Agentic AI’ is the natural evolution of artificial intelligence. They define it as an unavoidable transformative force for the business fabric, because it allows autonomous systems to discover, negotiate and coordinate data across the entire Internet infrastructure. It represents an unprecedented opportunity to drive productivity and innovation, but also presents complex ethics, safety and talent challenges.
For more information and a practical guide on how to meet this challenge, SAS has recently published an eBook, “A practical guide to understanding a new era of Artificial Intelligence.” The manual delves into these concepts, clearly differentiating Agentic AI from generative and general. In addition, it provides the essential keys and tools for its productive and governed implementation.
