Tech | Visa | Scholarship/School | Info Place

Google Cloud aims to simplify agent creation with Vertex AI Agent Builder

Artificial intelligence agents are the new craze in the field of generative artificial intelligence. Unlike previous generation chatbots, these agents can do more than just answer questions. They can take action based on conversations and even interact with back-end transaction systems to take action in an automated manner.

On Tuesday at the Google Cloud Next conference, the company unveiled a new tool to help companies build artificial intelligence agents.

“Vertex AI Agent Builder enables people to build conversational agents very easily and quickly,” said Thomas Kurian, CEO of Google Cloud. “You can build and deploy production-ready, AI-driven generative conversational agents with Instruct them the same way humans do to improve the quality and correctness of model answers.”

The no-code product builds on Google’s previously released Vertex AI search and conversation product. It is also built on the company’s latest Gemini large language model and relies on the RAG API and vector search, two popular methods used industry-wide to reduce hallucinations, when a model fails to find an accurate response. The model will give the wrong answer.

Part of the way the company improves the quality of its answers is through a process called “grounding,” in which answers are linked to sources considered reliable. In this case, it relies on a Google search (which may or may not actually be accurate).

“We are now giving you the foundation of Google Search, bringing the power of the world’s knowledge that Google Search provides through our foundational services to models. In addition, we are supporting the ability to target enterprise data sources,” Kurian said. The latter may be more suitable for enterprise customers.

In a demo, the company leveraged this feature to create an agency that could analyze previous marketing campaigns to understand the company’s brand style and then apply that knowledge to help generate new ideas consistent with that style. The demonstration analyzed more than 3,000 branded images, descriptions, videos and documents stored on Google Drive related to the fictional company’s products. It then helps generate images, captions and other content based on its understanding of the fictional company’s style.

While you can build any type of agent, this particular example would put Google in direct competition with Adobe, which released idea generation AI tool Firefly last year and GenStudio last month, to help build without straying too far from that. Corporate style content. There’s flexibility in building anything, but the question is whether you want to buy something off-the-shelf (if any).

According to Google, the new feature is already available.It supports multiple languages ​​and provides country-based API endpoints in the US and EU

#Google #Cloud #aims #simplify #agent #creation #Vertex #Agent #Builder

Leave a Reply

Your email address will not be published. Required fields are marked *