Artificial intelligence has ceased to be merely a support tool and has become a strategic partner for companies and professionals. The popularization of generative models was the gateway: now, we are experiencing the rise of Agency AIs, systems capable of operating autonomously, with decisions and executions in real time.
More than just agents restricted to specific tasks, these models take over end-to-end processes, from data analysis to decision-making, reducing costs, accelerating operations, and increasing efficiency in different sectors. In the financial market, for example, they are already applied to analyze large volumes of data, detect fraud, and anticipate risks, protecting investments and clients. In retail, they function as invisible inventory managers, predicting demand peaks, automating purchases, and adjusting prices according to consumption patterns.
The relevance is confirmed by market projections. According to the reports "Emerging Tech: The Future of Agentic AI in Enterprise Applications" and "Emerging Tech: Maximize Opportunities While Managing Risks of Agentic AI on Enterprise Software," both from Gartner, approximately 30% of global enterprise software revenue could be driven by Agentic AI by 2035, representing more than US$450 billion.
Companies as the gyms of the future.
This evolution brings a new requirement: preparing professionals capable of building, implementing, and governing robust AI projects. This is an advanced technology that demands skills far beyond basic or superficial knowledge. Understanding its true functionalities requires continuous, applied, and practice-oriented learning. In this scenario, companies cannot limit themselves to adopting tools; they need to assume the role of true learning centers, capable of developing talent and sustaining innovation.
This need becomes even more evident when we observe that many professionals are unable to solve systemic challenges precisely because they do not understand how certain knowledge applies to their daily work. This occurs because, to a large extent, they still rely solely on theoretical or self-taught learning – paths that, while valid, are insufficient in the face of current technological complexity. The study "Work: In Progress – Discoveries of how AI is transforming work," in which Google interviewed more than 3,500 professionals from different sectors in Brazil and other countries, illustrates this reality well by showing that 47% of Brazilians seek to learn about AI on the internet.
This learning model is relevant, but it alone does not sustain the full potential of Agency AI. It is necessary to go further, applying knowledge critically and contextually, expanding learning to aspects that go beyond technique and engage with the reality of organizations. This includes ethical concerns, such as the evaluation of biased systems, and also governance issues, such as defining responsibilities for decisions made autonomously by algorithms.
Organizations need to establish themselves as vibrant learning environments. This means investing not only in formal courses, but also in continuous upskilling and reskilling programs, mentoring, and practical experience in large projects. Learning in teams, side-by-side with peers and more experienced professionals, is what transforms knowledge into applied competence.
In the era of Agency AI and accelerated change, acquiring new skills will never be a destination, but rather a constant path to transform innovation into a lasting competitive advantage. Each new learning experience and application also represents an opportunity to generate value for the business.
*Guilherme Pereira is the Innovation Director at Alura + FIAP Para Empresas , corporate education solutions from the Alun Group, the largest business and technology education ecosystem in Latin America, supporting organizations of all sizes and segments in career and business transformation.
