AI嵌入旧流程难释放价值，在巴中资企业需重构业务架构

Silvio Meira Keiny Andrade/Valor Companies are still unable to realize artificial intelligence’s full potential because they are trying to fit it into outdated processes without reconsidering their business architecture or the role of people within their organizations. That is the assessment of Silvio Meira, chief scientist at TDS Company, co-founder of Porto Digital, and one of Brazil’s leading experts on the relationship between AI and the future of work. Meira advocates restructuring companies and changing the role of human beings, who should become orchestrators of intelligence and validators of the solutions produced by machines. In a conversation with Elisa Campos, editor-in-chief of Época Negócios, during the 2026 Valor Innovation Brazil Awards in São Paulo earlier this week, Meira described the transformation caused by artificial intelligence as the greatest ever produced by a general-purpose technology. “That is why, in this era of computing, if we do not redesign processes and understand the role of professionals as orchestrators of intelligence and architects who interpret what AI has created, we will not be able to add as much value as we should.” Below are the main excerpts from his interview with Valor. Valor: You often cite a concept from Peter Drucker [regarded as the father of modern management]: “The ultimate purpose of innovation is not to create a product or process. It is survival.” Could you explain that idea? Silvio Meira: We sometimes delude ourselves into thinking that innovation is equivalent to achieving better performance, higher quality, lower costs, and greater value creation to increase margins and profitability. But innovation, as a continuous process, occurs when businesses anticipate changes in their markets—particularly during periods of more radical innovation. Perhaps the most catastrophic example of the past 25 years was the iPhone’s arrival in 2007 in what appeared to be a cellphone market. At the time, Nokia invested more in research, development, and innovation than Apple’s total revenue. Steve Jobs’s company was not even entering the same market as Nokia. Yet three years later, the cellphone business was dead, replaced by the global smartphone market. Why? Because Nokia failed to anticipate the market’s transformation. Transformational innovation requires a different process for interpreting reality. It is not a present that carries some version of the past into the future. It is a present that brings new futures into the past. That is how innovation becomes the art of surviving change. Valor: What practical advice would you offer leaders trying to balance the need to deliver strong short-term results with the imperative to innovate and secure their companies’ future? Meira: Whenever someone asks me for the best solution to a complex short-term problem, I tell them it is long-term education. How do you solve the innovation problem in the short term? You cannot. Innovation is a long-term challenge tied to an organization’s incentive structures. If those structures reward only short-term results, and executives remain at companies for only two or three years, the vast majority of attempts at structural innovation will fail because senior leaders will not be involved. Innovation involves considerable risk and a significant possibility of failure. If I am not recognized for trying, making mistakes, and learning from failure—or if I am often punished for doing so—why would I try? Structural innovation within companies therefore requires courage. We need the courage to choose the problems and decide where to innovate. There is a century’s worth of literature on corporate change, which was not previously called innovation. Companies generally begin innovating at the periphery, not at the core of the business. They identify an opportunity on the margins and treat it as a laboratory for experimentation and performance. They test concepts and hypotheses to determine whether they can create a prototype or pilot that can be scaled. But when the time comes to implement it, they struggle to defend the change because they know they are putting their corporate futures at risk. Championing change is always difficult because everything within a business is designed to preserve stability. Valor: What innovation practices or models should Brazilian companies reassess? Meira: The vast majority of companies, regardless of type, do not have a strategy. They merely have a list of aspirations. The difference is that strategy is a continuous, deliberate, collective process that turns aspirations into achievements. If an organization has no strategy, it can put any wish or aspiration into its plan because none of it is binding anyway. What I see today, therefore, is very little strategy and a great deal of talk. Everything is motivational. This creates a kind of innovation theater in which people gather to discuss change and hold events intended to bring about change, but ultimately very little changes. Valor: If you became CEO of a large Brazilian company today, what would your priorities be? Meira: First, I would address the context in which the company operates honestly and realistically. Today, many leaders operate in an imaginary country that boards of directors commonly call Narnia. But the truth is that you issue invoices and pay bills right here in Brazil. For example, for a long time I heard people say interest rates were going to fall without any evidence to support it. Living in your imagination does not solve anything. Second, I would assess the organization’s capabilities honestly and realistically. Finally, leaders need to understand Ashby’s law [the law of requisite variety, developed by W. Ross Ashby in the 1960s, holds that for a control system or organization to maintain stability, its complexity and capacity to respond must be at least equal to the complexity of the external challenges it faces]. It means that to address a complex problem, an organization must understand and command that problem’s complexity. In innovation and marketing—two strategic areas—a company must know precisely what needs to be done, even if it hires an outside firm to execute the work. Otherwise, it will not even know whom to hire. Returning to my earlier point, innovation is risky. Anyone unwilling to take risks will continue doing exactly what everyone else is already doing. Every major company in the world today took a risk, recognized that a new market existed and that it might be possible to build a business around it, and invested in that possibility. Valor: There is now a consensus that AI is comparable to major technological revolutions such as electricity or computing. Even so, companies still struggle to extract value from it. What are they doing wrong? Meira: Companies are not changing their processes and business architecture enough. They are not leaving old practices behind to learn new things and build from them. Let me give you an example. What did steam-powered factories do when electric motors emerged in 1880? They replaced the steam engine with one large electric motor and left the entire factory unchanged, still powered by transmission belts and chains. The processes and productivity remained the same. It took 40 years, from 1880 to 1920, for everyone to understand that they needed lighter machinery powered by smaller motors distributed throughout the factory. Energy was no longer being distributed through mechanical drives but through electric current. That fundamental transformation wiped out almost every company that relied on steam power. The same process has occurred dozens of times throughout history, yet we never remember it. Today, artificial intelligence can perform a large share of the repetitive cognitive work humans do. In my field, for example, 95% of all code previously written by humans can now be written just as well—or better—by AI. Given that, what is the human role? A machine can write tens of thousands of lines of code a day, but who will read those lines to determine whether they do what they are supposed to do, whether they violate any ethical, moral, or legal principle, and whether they fit the market envisioned at the outset? Humans will. That is why, in this era of computing, if we do not redesign processes and understand the role of professionals as creators of context, orchestrators of intelligence, and architects who interpret what AI has created, we will not be able to add as much value as we should. Keeping human beings in the same positions, doing the same things only faster, does not create the value it should. Valor: What is the most interesting thing AI has to offer? Meira: Throughout human history, we have used two processes to intervene in reality: abstracting problems into questions and turning answers and solutions into reality. We surround a problem with questions to encode it, creating a kind of specification. We then use our existing knowledge to transform the problem into questions and answers. To turn the answers into reality, we use innovation, abstraction, and knowledge. Everything we had created until now was intended to produce things that did not compete with our brains. We created chairs so we could rest, buildings to shelter us, microphones to capture our voices, glasses to help us see near or far, and containers to carry all kinds of things. But we had never created anything capable of competing with our ability to abstract and turn ideas into reality. The question-and-answer issue was already solved in 2023 with the popularization of language models. Today, AI systems are doing very different things, such as solving a mathematical problem that had remained unsolved for 80 years. We are talking about an achievement that would once have earned a mathematician a Nobel Prize. The challenge for all of us—leaders and those they lead—is now to live in this new world, where information systems in front of us can imitate a very large share of humanity’s repetitive cognitive capacity at speeds no human can match. Translation: Todd Harkin