Artificial intelligence has entered a new phase. It is no longer merely a reactive tool but increasingly an autonomous agent, capable of defining objectives, making decisions and influencing human choices. This was the starting point of the article published by ECO on the event “Agentes da Mudança – Agentic AI and the Future of Decision-Making”, organised by WhatNext.Law, a research centre resulting from the partnership between VdA and NOVA School of Law, together with Capgemini, which brought together academics, businesses, regulators and public decision-makers in Lisbon.
The question posed by Margarida Lima Rego, from NOVA School of Law — “Who makes the decision when the agent is artificial?” — set the tone for the debate reported by ECO, focusing on the technological, legal, organisational and ethical implications of so-called Agentic AI. Throughout the event, it became clear that this is no longer a discussion about science fiction scenarios, but about technologies already being deployed, with AI agents operating in areas such as education, banking, insurance, telecommunications and critical infrastructure.
According to ECO, Cristina Rodrigues and Massimo Ippoliti, from Capgemini, stressed that the main challenge lies not only in the technology itself, but in data maturity, governance frameworks and mechanisms for control and audit. Miguel Mira da Silva, professor at Instituto Superior Técnico, reinforced the idea that AI initiatives should be treated as innovation projects rather than traditional IT projects, warning of the cultural and structural limitations still present in many organisations.
The ECO article also highlights several concrete adoption cases presented by representatives from BPI, Fidelidade, MEO and Brisa, illustrating how AI agents are already transforming internal processes, customer service and operational decision-making, while at the same time raising new requirements in terms of literacy, management models and trust.
From a legal and regulatory perspective, the article gives particular prominence to the discussion around the AI Act, with contributions from Raquel Brízida Castro, Vice-President of ANACOM, Vera Lúcia Raposo, professor at NOVA School of Law, and Magda Cocco, partner at VdA. The debate underscored how agentification forces a reassessment of the way risks, responsibilities and governance models are mapped, in a context where data has definitively moved beyond a compliance issue to become a core governance concern.
The ECO piece concludes with a broader reflection on trust, AI literacy and the role of the State, with contributions from Arlindo de Oliveira, Joana Gonçalves de Sá, Paulo Dimas and the Secretary of State for Digitalisation, Bernardo Correia, reinforcing the idea that Agentic AI represents not only a technological evolution, but a structural test of how society, organisations and regulators choose to make decisions in an increasingly hybrid future of humans and artificial agents.
Read the full article published by ECO here
