Google's AI agents will completely change the way we interact on the web

Google seems to be putting focus on AI agents. This technology is starting to generate good impressions due to the fact that the user could automate their tasks in a really amazing way. There has been a lot of talk at the Google I/O conference. But again, everything is focusing on AI agents, which could change the whole world as we know it today.

You may have noticed some changes when browsing and searching for information. This change is influenced by the implementation of Gemini in Google's search engines, which offers brief information through the AI-created Overview. These answers are adapted with respect to various sources, making a summary of what was searched, although there was much talk about the hallucinations that the AI could generate, so it was something that should be polished. Also something that could be seen is the implementation of Gemini in other Google applications, so the idea of embracing everything goes in that direction.

Now, talking about other technologies, it is logical to think that there will be a communication between them, so that the product offered manages to captivate the user, and potential customer. If there is something to say, it is that AI agents reflect a perfect intermediary when interacting with the web. The words of Google's Vice President of Search, Liz Reid, generate greater enthusiasm for this technology by saying that AI will be a catalyst for potential discovery never before seen on the web. In addition, it will be possible to request information through complex questions or research, and even access personalized purchases.

One of the announcements that transcended at the event was the AI mode, available to all users residing in the United States. This feature allows you to interact through an AI agent, either to take care of visiting websites, making summaries, or making purchases, which to my mind is a feature that improves productivity because of how these agents work; in a timely and accurate manner. On the other hand, a subscription to Ultra will allow the user to access Project Mariner; an agent capable of simultaneously performing up to 10 tasks, giving the user flexibility and time to do other things.

---

Obtained from Google's blog - Google AI Search AI Mode

---

Also unveiled was the progress of Project Astra, which would not benefit if it were not aligned with the development of Gemini Live, which allows the AI to see and hear. This is a multimodal AI that can understand and interact with its environment in real time, which generates a wide range of possibilities in our lives, having a great assistant available at all times. Knowing that they are working on this project, it is likely that we will see news about it in the following months.

The idea that they are creating an infrastructure that supports the connection of agents to data sources is incredible. Why? Because by becoming a standard, it will enable collaboration between AI companies, which would broadly benefit us. One case we should mention is the implementation of Google's Agent2Agent protocol and Anthropic's MCP, resulting in the interconnection between agents and web-based data sources.

However, unresolved issues have yet to be resolved. For example, the use of agents could cause a rupture in the way advertising is done on the Internet. Another issue to be questioned is that AI summaries could affect the way publishers do their work, having to compete against these agents. However, there is an unintended effect, which is that if these publishers lose their jobs, the AI will not be able to supply itself with data with which to train.

This and much more are issues to contemplate with respect to the growth and full expansion of these AI agents and models. It should not be a process of adaptation as such, but rather one in which problems are solved at the lowest possible cost.