What if the Future Is Not about Apps — but Agents that Think, Act, and Work Together?
This year, Google Cloud Next ’25 took place during the first week of April in Las Vegas, USA. The event lasted 3 days, featuring 10 keynotes and 700 sessions around innovation topics such as artificial intelligence, data cloud, modern infrastructure, security, Google Workspace, among others.
And I was there to hear firsthand about the main trends coming for this new year! In this article, I would like to share a few thoughts on a new development that is likely to spark conversation in the near future.
From SaaS to Agents-as-a-Service: What Is Next in Digital Product Development according to Google Next ‘25
At this event, Google did not just display new products — it put forward a new perspective on software. The trends made it clear that we are entering the “agentic era,” where digital product design is shaped by intelligence, autonomy, and collaboration between agents.
This novel approach intends to replace traditional apps with agent ecosystems that, powered by large language models (LLMs) and modular architectures, are capable of reasoning, planning, and executing complex tasks in real time. At the same time, they can swiftly adapt to the changing needs of users and companies.
This concept of Agents-as-a-Service (AaaS) represents a significant evolution from the SaaS (Software-as-a-Service) model. While SaaS focuses on providing access to apps, AaaS goes a step further by offering intelligent agents that can interact not only with data and systems but also with each other. This communication between agents is enabled by protocols like Agent2Agent, which allow for seamless and efficient collaboration toward shared goals.
So, what would exactly change with this approach?
- The information model
Digital products are no longer just about organizing information and helping users visualize it. Agents process natural language, understand the context, and are capable of making decisions over dynamic flows. This implies designing user experiences where people are not only entering data and clicking, but also initiating conversations, suggesting ideas, and awaiting responses.
Thus, product design shifts from defining static flows to curating user behavior, especially as we now work with systems that learn, evolve, and react.
2. Teams’ skill sets
Adapting to this era means teams need to develop new skills such as conversational experience design, agent creation and fine-tuning, definition of trust and explainability metrics, orchestration of multi-agent flows, cybersecurity, data analysis and processing, among others.
3. The ability to understand the problem as a new strategic asset
AI architecture and agent architecture both require clear instructions, and to generate those, one must deeply understand the problem to be solved. But that problem is no longer static — it evolves along with data, users, and the business itself. Problem framing and interdisciplinary teamwork will be essential to tackle this challenge, where breaking down the traditional lines between technical and non-technical roles is necessary.
4. The definition of a digital product
Digital products have evolved from closed flows into dynamic ecosystems where various agents work together to support user needs. In a context with fewer hardcoded rules and more emergent, unpredictable behavior, the focus must be on enabling the system to learn from its users, rather than forcing users to adapt to the system.
5. Organizational autonomy
Agents make it possible to automate many repetitive decisions, but they also open the door to more decentralized models where each team or department can build its own flow with embedded intelligence. This frees central areas from having to orchestrate everything and enables scalability, which also brings along a massive challenge in terms of organizational change, decentralized governance, and security.
So, what are the benefits?
- Reduced operational time: Tasks that used to take days can now be completed in minutes thanks to agents that interpret documents, analyze the context, and provide recommendations.
- Improved user experience: Instead of requiring users to understand complex processes or adapt to rigid rules, agents adapt to the users. This enables smoother, more accessible, and inclusive experiences that do not revolve around navigating a structure but rather achieving a goal with intelligent support.
- Scalability: Agents allow intelligence to be distributed in a modular and comprehensive way across organizations. Rather than relying on centralized systems or rigid hierarchical structures, each unit or team can have agents tailored to their own processes, integrated with their data and goals. This decentralization enables more organic scalability, where each part of the system can evolve, learn, and improve without requiring a global redesign.
A perspective from Flux IT
I see this stage as a new logic that changes the way we create value. However, it is important to identify when agents can truly add value, for instance, in scenarios involving high-volume data processing or customization, for repetitive decision-making, costly human intervention, or complex flows with many interactions.
But even if not all processes need agents, many do need to stop being rigid, impersonal, or inefficient.
So, what is the key? To focus on people and help organizations build digital solutions that not only work but understand, support, and adapt to their context. Because with this perspective, we are getting closer every day to achieving that products learn from people — not the other way around.
