Do use cases still make sense in the age of AI agents?
Raphaël Fétique, President and co-founder of Converteo, has been a data and digital expert for over 15 years. A graduate of Télécom Paris and HEC, he co-founded Converteo in 2007 with Thomas Faivre Duboz.
In a frantic race to experiment, companies are piling up AI “use cases,” hoping to find a decisive advantage. This fragmented approach, a legacy of twenty years of digital transformation consulting, blinds us to the true nature of the ongoing revolution: the rise of autonomous agents, which demands a systemic shift in vision rather than a collection of projects.
Use case: the illusion of mastering AI?
The scene repeats itself in every boardroom. On the screen, a list of potential “use cases” for artificial intelligence (AI), ranked by complexity and expected return. The discussion revolves around optimizing a process, predicting sales, or automating a recurring task. The exercise is familiar, almost comforting. It creates the illusion of mastering a complex technology by breaking it down into digestible problems, each promising a quantifiable return on investment. This method proved successful for digital, so why wouldn’t it work for AI?
The reality is that this approach is leading us to a strategic dead end. By focusing on these one-off applications, we are missing the bigger picture.
The comforting tyranny of the application
This obsession with use cases is no accident. It is the direct legacy of the methods that have governed digital transformation since the beginning of the 21st century. To tame the complexity of the internet, software, and platforms, we learned to isolate perimeters, define projects, and measure short-term results. The “use case” became the alpha and omega of innovation, the elementary particle of every strategic plan. It reassures CFOs and structures the work of teams.
Applied to the first wave of AI—that of predictive models—this logic still worked. A model was designed for a specific task: detecting fraud, recommending a product, anticipating a failure. The process was linear, its consequences contained. But the era of generative AI, and even more so the emerging era of agentics, shatters this paradigm.
From tool to agent: a revolution in thinking
The shift from AI-as-a-tool to AI-as-an-agent is a true conceptual rupture. Therefore, asking for the “use case” of an agent would be like asking for the “use case” of an intern or an employee: the question makes no sense.
You don’t hire an employee for a single, repetitive task, but for their ability to contribute to a broader mission by leveraging a variety of skills. The value of an agent lies not in a single application, but in its ability to orchestrate multiple actions to achieve a goal. This is a fundamental transformation: the focus shifts from the application of artificial intelligence to its integration into the core of workflows.
Thinking in capabilities, not projects
The challenge for leaders is no longer to collect proofs of concept. It is now about rethinking the organization of work and corporate culture to leverage these new non-human collaborators. The relevant question becomes: “what new strategic capabilities can we build thanks to hybrid teams of humans and agents?”.
This requires managerial courage: the courage to move from a logic of control-by-project to a logic of trust and delegation to the machine. It requires investing less in isolated technical solutions and more in team acculturation, fundamental data quality, and the redefinition of business processes. It’s no longer about asking what AI can do for a job, but how a job can reinvent itself with AI.
The time for scattered efforts is over. The companies that succeed will be those that stop chasing technological butterflies and focus on building an ecosystem where intelligent agents amplify human capabilities. They will have understood that artificial intelligence is less a technological issue than a profound cultural transformation project.
