Vertex AI Agent Builder is the tinkering corner for generative AI on Google Cloud. The platform enables companies to develop and deploy self-created AI agents.

Google Cloud NXT launched today in Las Vegas. It should surprise no one that (generative) AI is a main topic of the conference. Despite the sometimes bumpy path the company has taken over the past year, Google wants to convince the world that it is the AI ​​specialist par excellence. This happens not only with ready-made generative AI, but also by giving companies the necessary tools to create applications themselves.

With Vertex AI Agent Builder, Google Cloud opens a new corner for AI enthusiasts. Vertex AI is a collection of cloud services that companies use to develop machine learning models. Through Vertex, companies gain access to AI models from Google and recognized partners. They can then train these models using their own data, developing “AI agents” that are specialized in the defined business context.

Craft without code

The Vertex AI Agent Builder is designed to make this entire process as simple and accessible as possible for Google Cloud customers. The tool works without code Natural language interface, so no code or programming knowledge is required. The platform allows developers to create new agents in minutes by defining goals in natural language, providing step-by-step instructions, and sharing conversation examples.

These agents can connect to corporate data to improve response accuracy, invoke functions, or connect to applications to perform tasks for the user. For complex purposes, the Vertex AI Agent Builder also allows multiple agents to be linked. The platform includes access control, data management, and data sovereignty options, allowing developers to tinker while keeping data secure.

Based on data

An AI agent works in a defined context, so Vertex AI Agent Builder provides features to anchor models in business data. Vertex AI Search is a RAG system ((Augmented Generation Retrieval) This allows AI agents to gain controlled access to databases outside of their training sets to increase the accuracy of the output. Through Vertex, companies can also connect their models to Google Search to get up-to-date data.