The money needs to reach companies quickly, or the entire strategy stalls.
Portugal's Recovery and Resilience Plan, conceived before the generative AI era, has quietly repositioned artificial intelligence as the cornerstone of its national modernization strategy. Across factory floors in Águeda and government offices in Lisbon, the technology is already reshaping how work is done — yet the country's ambitions outpace its execution, with promised funds delayed and digital skills unevenly distributed. The deeper question Portugal now faces is not whether to embrace AI, but whether its institutions can move with the urgency the moment demands.
- A plan written in 2021 had to reinvent itself mid-course as AI transformed from distant trend to industrial necessity, forcing Portugal to bet its modernization on technology it never originally planned for.
- A €200–300 million instrument targeting over 2,000 SME projects signals serious public commitment, but companies that hired teams and launched projects in good faith are still waiting for advance payments that haven't arrived.
- Traditional manufacturers — textile firms, factories — are already logging real productivity gains from AI-driven systems, proving the technology works on the ground even as the funding machinery stalls above them.
- Digital literacy remains the quiet crisis beneath the headline numbers: without systemic upskilling across schools, universities, and workplaces, the workforce risks being left behind by the very transformation the plan is funding.
- Portugal's oversight commission chair has acknowledged a 'disconnect between public discourse and actual execution,' framing faster disbursement and nationwide skills investment as the two levers that will determine whether the window of opportunity stays open.
Portugal's Recovery and Resilience Plan was never designed with artificial intelligence in mind. Drafted in 2021, it predated the generative AI wave entirely. But by 2024, the technology had moved so fast that the plan had to move with it — and today AI stands at the center of Portugal's industrial and governmental modernization, not as an addition but as its most visible engine.
The evidence on the ground is concrete. At a factory in Águeda, sensors feed real-time data into systems making near-instant production decisions. At a textile manufacturer, tablets at work posts allow production lines to reorganize fluidly. These are not tech startups — they are the traditional firms that form the backbone of the Portuguese economy, and they are finding that AI solves real problems. The state, too, has begun using automation to process grant applications and analyze documents, always with human oversight, always with the aim of cutting bureaucracy.
To accelerate adoption, the government launched a financial instrument in late 2025 worth between 200 and 300 million euros, targeting more than two thousand SME projects across sectors from pharmacies to professional services. The ambition is that growing domestic demand will in turn expand Portugal's own AI startup ecosystem.
Yet two structural problems keep surfacing. The first is payment delays: companies approved for support, having already hired staff and launched projects, have not received the advance funding they were promised. The oversight commission chair acknowledged the gap between public rhetoric and actual execution, and said he had been pressing the Portuguese Development Bank to move faster. The second problem is digital literacy — a foundational deficit that no single program has yet addressed at scale. One retraining initiative, Upskill, demonstrated that redirecting professionals from fields like nursing or design into technology roles could pay for itself within three years. But such programs remain exceptions.
The Recovery Plan has been reshaped repeatedly by pandemic, inflation, war, and labor shortages, each crisis nudging it further toward direct private-sector support and deep technology. What has emerged is a country with the pieces in place — growing innovation hubs in Lisbon, Porto, Braga, and Coimbra, a clearer strategic vision, and real early results — but one that must now prove it can disburse funds quickly, build skills broadly, and sustain its direction long enough for the investment to compound.
Portugal's Recovery and Resilience Plan arrived at artificial intelligence late, almost by accident. The framework was drawn up in 2021, before generative AI had seized the world's attention, before ChatGPT existed, before anyone could have predicted how quickly the technology would move from laboratory curiosity to factory floor necessity. But by 2024, the landscape had shifted so completely that the plan itself had to shift with it. Today, AI sits at the center of Portugal's modernization strategy—not as an afterthought, but as one of the most visible engines driving both industrial productivity and the digitalization of government.
Pedro Dominguinhos, who chairs the oversight commission for the Recovery Plan, traced this transformation in a recent podcast conversation. He described a country that had once regarded artificial intelligence as a distant trend, something for Silicon Valley and Beijing to worry about. Now it was woven into production lines, embedded in business decision systems, integrated into how the state processed applications and managed its operations. The shift happened fast, and the evidence was concrete. At Miranda & Irmãos, a factory in Águeda, sensors distributed across the shop floor fed real-time data into systems that made production and logistics decisions almost instantaneously. At Calvelex, a textile manufacturer, tablets stationed at work posts helped reorganize production lines whenever operational changes occurred. These were not tech companies experimenting with cutting-edge tools. These were traditional industrial firms, the backbone of the Portuguese economy, discovering that AI could solve actual problems.
The transformation extended beyond factory gates. Portugal's public administration began deploying AI for document analysis and application processing. The Gov.pt platform incorporated AI-powered features. The Recuperar Portugal structure used automation to review grant applications, always with human oversight, always with the goal of cutting bureaucracy and accelerating decisions. The state was learning to move faster.
To accelerate this shift, the government launched a financial instrument in late 2025 that would distribute between 200 and 300 million euros to small and medium-sized enterprises adopting AI solutions. The program aimed to support more than two thousand projects across sectors as varied as pharmacies, manufacturing, marketing, and professional services. Dominguinhos argued that the impact would ripple outward. As demand for customized solutions grew, Portuguese startups and domestic technology providers specializing in AI would find new markets. The ecosystem would expand, creating opportunities for everyone inside it.
Yet beneath the optimism lay a structural problem that kept surfacing in conversation: digital literacy. Portugal lacked the foundational skills its workforce needed to compete in an AI-driven economy. Dominguinhos argued that the country needed to treat AI literacy as genuine national infrastructure, something that involved schools, universities, technical institutes, private companies, and government agencies working in concert. He pointed to the Upskill program, developed through partnerships between tech firms, academia, and the national employment service, which had retrained professionals from psychology, design, and nursing into information technology roles. The program paid for itself within three years through higher salaries, retained talent, and increased tax revenue. But such initiatives remained exceptions, not the rule.
The conversation also exposed a gap between rhetoric and reality. Elsa Veloso, one of the podcast hosts, raised the issue of payment delays. Companies approved for AI-related support programs, particularly through the Financial Instrument for Innovation and Competitiveness, had not received the advance payments they were promised. Many had already hired teams and launched projects on the assumption that funding would arrive. It hadn't. Dominguinhos acknowledged the problem directly, admitting there was a "disconnect between public discourse and actual execution." He said he had been pressing the Portuguese Development Bank to accelerate payments. "It's essential that the money reaches companies quickly," he said.
The Recovery Plan itself had evolved considerably since its inception. Pandemic, inflation, energy crisis, war, labor shortages—each had forced adjustments. The result was a rebalancing toward direct support for private enterprise, with particular emphasis on deep technology, defense, and artificial intelligence. Dominguinhos argued that Portugal now needed to create conditions for startups to become global scale-ups, strengthening venture capital mechanisms, international connections, and market linkages. He noted that Lisbon, Porto, Braga, and Coimbra had begun to feel more international, more connected to global innovation networks.
The central message that emerged was this: artificial intelligence had stopped being merely a technological question. It had become a strategic tool for economic competitiveness. Portugal had the pieces in place to build a more robust innovation ecosystem—but only if it could accelerate the disbursement of funds, raise digital skills across the population, and maintain a consistent long-term vision. The window was open. Whether the country could move fast enough to walk through it remained the question.
Citações Notáveis
There is a disconnect between public discourse and actual execution— Pedro Dominguinhos, chair of the Recovery Plan oversight commission
It is essential that the money reaches companies quickly— Pedro Dominguinhos
A Conversa do Hearth Outra perspectiva sobre a história
Why did AI take so long to enter Portugal's Recovery Plan in the first place?
The plan was designed in 2021, before generative AI had exploded into public consciousness. It was built around different priorities. But by 2024, the technology had become so obviously useful—so immediately applicable to real production problems—that the plan had to adapt. That's actually what happened.
So this isn't a failure of foresight, it's a success of flexibility?
Partly. The plan was flexible enough to shift. But it also reveals something about how slowly institutions move. By the time AI became undeniable, companies were already experimenting on their own. The government was playing catch-up.
The funding sounds generous—200 to 300 million euros. Why are companies still waiting for payments?
That's the real problem. The money exists on paper. The programs are approved. But the actual disbursement—the moment when cash hits a company's account—is stuck in bureaucratic processes. Companies have already hired people and started work. They're operating on faith that the money will arrive.
What happens if it doesn't?
They fail. Or they slow down. Or they become cautious about hiring. The entire ecosystem stalls. That's why Dominguinhos kept saying the money needs to move faster. It's not just about fairness—it's about whether the whole strategy actually works.
He mentioned digital literacy as a major obstacle. How serious is that?
Very. You can fund AI adoption all you want, but if your workforce doesn't understand how to use these tools, or how to think about them, the money is wasted. The Upskill program showed it's possible to retrain people quickly—three years to payoff—but that's one program. You'd need this at scale across the entire country.
Is Portugal actually building that scale?
Not yet. That's the gap between the ambition and the reality. The vision is clear. The execution is still catching up.