It has never been easier to experiment with Artificial Intelligence. A manager can open a generative AI tool, automate a task, or prototype an agent in an afternoon. The challenge is no longer access to AI. It is turning that experimentation into organizational capability and, over time, into a stronger business model.
Experimenting and transforming are different things. One team automates reporting; another builds a customer assistant and another tests AI agents. Each initiative may create value, but their sum is not, in itself, an AI strategy.
AI adoption in organizations tends to evolve in stages. First, it transforms tasks: analysis becomes faster, content cheaper, and administrative work easier. Next, it transforms workflows: tasks are recombined, responsibilities shift, and people increasingly supervise processes partly executed by machines. The third stage is the strategic one: AI changes how customers search, choose and buy and, with that, the business model can begin to change.
Retail is a good example. For decades, companies such as Amazon gained strength by bringing together discovery, comparison, transaction and fulfillment. AI agents can separate these activities again. Imagine a consumer asking an AI agent: “Find me the best running shoes under €150, compare the options and buy the best pair.” The retailer still holds the product and handles delivery, but another player starts to influence the most valuable moment: deciding what to buy. This already carries economic weight: AI-referred traffic to retailers is growing rapidly, and Amazon itself is already investing heavily in agentic shopping.
The question goes beyond internal use within companies: leaders need to understand how AI could change the relationship with customers, which activities become commodities, and where competitive advantage will sit.
Scaling AI therefore requires answering three questions: Where can AI improve efficiency today? Which workflows should we redesign next? And which parts of the business model could AI transform tomorrow?
This is where the conversation shifts from technology to strategy. Leaders have to decide which capabilities they want to keep in-house and which to “buy”, where human judgment remains indispensable, and which positions in their ecosystem they cannot afford to lose.
Skills matter too. Knowing how to write prompts is no longer enough: people need to know when to trust AI, when to challenge it, and how to combine it with human judgment. As Academic Director of The Lisbon MBA Católica|Nova, I see this reflected in management education: AI should serve not only to increase productivity, but to challenge students, as leaders, to question results, compare alternatives and better justify their decisions. Understanding the technology is important, but their greater responsibility will be to redesign organizations and decide where AI should, and should not, transform the business.
This is also why the number of pilot projects is no longer a good measure of maturity without a clear strategy. Lasting advantage will not belong to the organizations with the most pilots, but to those that move deliberately from tools to processes, and from processes to business-model choices.
The tools will keep changing. The real strategic task is to understand what should change with them, and what should remain distinctively yours.
René Bohnsack
Full Professor of Strategy and Innovation at Católica-Lisbon SBE, Academic Director of The Lisbon MBA Católica|Nova and author of “The AI Strategy Playbook: From Zero to 100”
📖 Read the original article in Portuguese here.
