If you can imagine a plausible near future, imagine this situation: Your most important client has summoned you to an urgent meeting at his company. Your AI, your AI agent, which you have named Jarvis (I will not deny that I am a fan of science fiction), has reviewed your client’s appointment in the agenda, has calculated the best route from your position to the client’s company, has remotely activated your electric car so that it is heated when you pick it up, and has negotiated with another AI agent the best parking space in a multimodal hub close to the destination. This situation, which is fictitious but plausible in the short term, was the focus of the debate at the last edition of the Mobile World Congress (MWC) 2025 held in Barcelona.
Yes, it is true, AI has become a label for everything – proof of this has been the biggest event of the year held in our country – but under its umbrella there are signs that outweigh others and that promise to shake up everything we know. Agentive AI is one of these signals.
But what in the world is agentive AI? Well, to put it simply, it’s that assistant that not only writes an email on your behalf, but decides when to send it, to whom, and then schedules a follow-up meeting on your calendar. All while you’re watching the latest season of TV series “Black Mirror” (ironies of life).
This is not a trivial matter. In the past, the user chose, planned or booked. Today, AI predicts, suggests and optimizes. But tomorrow… our agent will interact, negotiate, book and make decisions without human intervention. It won’t need you. That’s why it’s important to imagine a future where you yourself take a back seat.
As Joe Atkinson, global head of AI at PwC explained during the congress, “AI agents sound scarier than they are.” They are AI systems that specialize in specific tasks that can interact with other agents to perform more complex work. “Think about having a calendar agent, one that just focuses on how my schedule is, and another agent that maybe manages my social networks. In the future, we’ll see the calendar agent coordinating with the social agent and letting a lot of that work happen for us instead of being done by us,” Atkinson added.
One of the most outstanding sessions of the congress was the one led by Pilar Manchón, Senior Director of Engineering, AI Research Strategy at Google. Manchón has more than 25 years of experience in the AI sector. She is a researcher at the universities of Seville, Edinburgh and Stanford, as well as an entrepreneur and executive at companies such as Intel, Amazon, Roku and now Google. Her participation in the MWC was a key factor. There, she explained that she is currently working on AI moonshots with conversational agents, that is, exploring ambitious projects that Google undertakes with no guarantee of short-term profit. During the session, Manchón defended a “human-centered artificial intelligence”, “of trust” and urged not to fear it but to “embrace it” to “change the world”.
The word trust takes on considerable relevance here. To embrace as she explained the services of these AI agents, we need humans to “trust the agents. They cannot be left to freewill or doing their own thing.” This is critical because we have no guarantees that everyone will develop agents in the best possible way and for the best possible purpose.
Will human beings continue to be the central stakeholder in marketing and communication strategies?
The question resonates more and more strongly. Until today, we had placed the conversation between the brand and the consumer at the center of the equation. People at the center, we used to proclaim, come hell or high water. But what happens when the relevant decision maker in the journey is no longer a person but an automated entity, an AI?
There are two potential scenarios: collaborative automation and pervasive automation.
Pilar Manchón likes the first one. She prefers trusted agents that cooperate with human beings. In collaborative automation the human decides, and the AI optimizes. There is a sense of control. And, therefore, trust. Here AI amplifies the capabilities of a human being in any of its roles, consumer, customer, employee, etc… The human being remains at the center of the strategy.
The other scenario is dark, dystopian, but likely. The AI makes decisions for you. You do not choose, the system decides for you, knows your patterns, anticipates and makes the best decision.
In this scenario the user just flows because the AI orchestrates everything.
Here the AI dominates the ecosystem. Humans only intervene – if at all – to validate, to get carried away. Human beings are displaced by the ecosystem and, therefore, disappear from the strategy to occupy a secondary role.
Does this mean that tomorrow we will design communication and marketing strategies to influence the decision making of AI and not humans? I leave the question there because it could be talked about long and hard.
The prompt economy
The MWC sessions did not end there. More meat was put on the grill, and a second fascinating term appeared: the Prompt Economy, a new paradigm where a simple instruction to AI drives entire workflows, unlocks experiences and achieves levels of efficiency previously unattainable for many companies.
Put another way: Imagine a world where you no longer pay for software. Nor do you pay for a subscription. You pay for the quality of your instructions, of your prompts. It’s as if instead of buying a car, you pay for the ability to tell it exactly where you want to go and how you want to get there. The difference between “take me home” and “take me home avoiding traffic, passing the gas station and playing my favorite playlist” might cost a few euros more, but the added value is exponential.
An economy that is redefining traditional business models, and where companies are developing new monetization systems based on pay-as-you-go services and dynamic billing systems.
And this is because in a world where the barrier between humans and machines is increasingly blurring, the ability to effectively “talk” to AI is fast becoming a crucial competitive advantage. An economic advantage.
