Considered one of the leading digital commerce events in Latin America, VTEX Day brings together leaders to discuss the future of retail, technology, and management. And at this year's event, one of the panels that provoked the most reflection was "AI Is Not Hype: It's a Competitive Advantage ," led by Patricia Florissi, PhD and technical director at Google.
The presentation wasn't just about technology. It was a lesson on how artificial intelligence has evolved and, most importantly, how it changes the way companies operate.
From understanding words to understanding the world.
To explain why AI is no longer just hype, Patricia went back to the technical basis that underpins everything we are seeing today.
The starting point lies in how machines have come to understand language. Instead of treating words in isolation, models have begun to organize them into mathematical structures, called "embeddings," capable of capturing meaning and context. This has allowed systems not only to identify words, but also to understand the relationships between them.
But the real leap came with the "Transformers" architecture, presented in one of the most influential papers of the last decade. From there, it became possible to analyze relationships between words within a complete context, with high processing power and scalability.
“This account showed the fundamentals of what happened in generative AI,” Patricia stated, demonstrating that the path had been opened for something bigger.
From there, the evolution was rapid. The same principle used for language began to be applied to images, videos, audio, and other data formats. The result is what is now called "multimodality," the ability of AI to understand different types of information in an integrated way.
The game-changing leap: AI that understands context and reality.
If the first phase of AI was understanding language, the next one is even more ambitious: understanding the world. The PhD highlighted the advancement of so-called multimodal models and, especially, " World Foundational Models ," which not only process data but can simulate and predict behaviors in the physical world.
In practice, this means moving from an AI that answers questions to an AI that anticipates scenarios. "We are empowering natural languages to have the same level of reasoning, in theory, as human beings," Patricia revealed.
This ability to integrate different sources of information (text, image, sound, sensors) and generate complex inferences is what underpins the next wave of business transformation. And this transformation, according to her, will not be gradual. "In the next 10 years, we will have an impact 10 times greater than an industrial revolution in one-tenth of the time," Patricia declared.
From e-commerce to "autonomous commerce"
To make this vision more tangible, Patricia Florissi presented a hypothetical, yet plausible, scenario of how consumption might evolve.
The story revolves around Monica, a character who uses a personal AI agent. This agent not only responds to commands, but also interprets context, anticipates needs, and executes actions autonomously.
Upon identifying a need (a non-existent work environment in a hotel, for example), the system:
- understands the context from multiple sources (agenda, sensors, environment),
- design an ideal solution,
- negotiates with suppliers,
- organizes logistics,
- and executes the delivery.
This example illustrates a structural change: consumption ceases to be reactive and becomes predictive and automated.
According to Patricia, this impacts the entire chain:
- the consumer becomes a generator of intentions,
- The salesperson becomes a solutions orchestrator.
- The product ceases to be fixed and becomes "liquid" (used on demand).
- and the buying journey becomes seamless and frictionless.
Models, agents and protocols: the new operational basis
For this scenario to work, Patricia organized the explanation into three technological pillars that are already under development: models, agents, and protocols.
Models function like the brain, responsible for understanding patterns, generating responses, and simulating scenarios.
The agents are the execution: goal-oriented systems capable of planning, acting, and replanning autonomously.
And protocols function like language, allowing different systems to communicate, negotiate, and operate together.
And here is one of the most profound changes for the business world: software is ceasing to be a passive tool and is beginning to act as an autonomous entity, capable of making decisions and executing tasks.
What does this change in practice for companies?
Despite the technical complexity, the impact on businesses is quite concrete.
We are moving towards an environment where:
- Decisions are made based on real-time context.
- Processes are automated from end to end.
- and operations are now coordinated by intelligent systems.
This has direct implications for efficiency, costs, and scalability, which are central themes for any financial area.
More than just adopting technology, the challenge becomes redesigning entire processes based on these new capabilities.
Automate or reinvent?
In the final part of the lecture, Patricia offered a reflection connecting technology and strategy, proposing a simple yet powerful distinction that is crucial to the role of leadership: should technology be used merely to automate the present, or to shape the future?
"Digital transformation is not about taking a snapshot of what already exists. It's about designing the future with technology," stated the PhD.
Based on this idea, she proposes the concept of "technographing," which would be the ability to imagine new scenarios and build them using available tools. And she poses a direct question: "Are we going to use AI to automate the present or to imagine and build the future?"
What remains for financial leaders?
The panel's idea reinforces that AI has moved beyond being a trend and has become a concrete lever for competitive advantage. For financial leaders, this means a change of role: moving from a focus solely on control to actively participating in the transformation of business models.
Efficiency, predictability, and control remain essential, but now within a much more dynamic, automated, and data-driven context.
And that's exactly where PagCorp positions itself: helping companies transform financial management into strategic intelligence, prepared for a scenario where technology, operations, and decision-making are becoming increasingly integrated.
Image: PagCorp
