Vertex AI Agent Builder lets developers, even those with limited machine learning skills, tap into the power of Google's foundation models, search expertise, and conversational AI technologies to create enterprise-grade generative AI applications.
Vertex AI Agent Builder is comprised of the following features:
- Vertex AI Agents
- Vertex AI Search
Vertex AI Agents
Vertex AI Agents is a new natural language understanding platform built on large language models (LLMs). It makes it easy to design and integrate a conversational user interface into your mobile app, web application, device, bot, interactive voice response system, and so on. Using Vertex AI Agents, you can provide new and engaging ways for users to interact with your product.
For more information, see Playbooks and Data store handlers in the Dialogflow CX documentation.
Vertex AI Search
Vertex AI Search is a fully-managed platform, powered by large language models, that lets you build AI-enabled search and recommendation experiences for your public or private websites or mobile applications.
Search apps
With Vertex AI Search, you can quickly build a Google-quality search app on your own data and embed a search bar in your web pages or app.
You can create the following different types of search apps:
Generic search. Apply generic search to websites or to data stores containing your proprietary data, giving your customers Google-quality search experiences on the content that you want them to see.
Media search. This is a search capability specially designed for media content such as movies, videos, and music. With media search, audiences can efficiently find the media content that they want to view or listen to.
Healthcare search. This is a search capability that lets you query healthcare records stored in FHIR data stores. You can import FHIR resources that contain clinical data from your Cloud Healthcare API FHIR store. You can also search unstructured data, such as images, PDF files, and RTF files, referenced by the FHIR resources.
Recommendation apps
You can quickly build a state-of-the-art recommendations app on your own data that can suggest content similar to the content that the user is viewing.
You can create the following two different types of recommendations apps:
Media recommendations. Get recommendations for media content such as videos, news, and music. With media recommendations, audiences can discover more personalized content, like what to watch or read next, with Google-quality results customized using optimization objectives.
Generic recommendations (Preview). Get recommendations for non-media content.
Key features
- Out-of-the-box natural language understanding and semantic search. Get a high-quality search experience without needing to implement and maintain systems that perform keyword searches or pattern matching.
- Out-of-the-box capabilities to understand synonyms, correct spellings, and auto-suggest searches. Improve the user's search experience without the need to implement complex natural language processing techniques.
- Generative AI. Get generative AI-powered summarization and conversational search for unstructured documents.
- Out-of-the-box recommendations. Get state-of-the art, ML-based content and metadata understanding that lets your users quickly find content that is similar to the content that they're viewing.
- Vertex AI Agent Builder console and APIs. Use the Agent Builder page of the console or Google's APIs to set up a search app or recommendations app for your public websites or for your structured or unstructured data.
- Out-of-the-box widget. Integrate search into your website. For more information, see Add the search widget to a web page.
- Self learning. Get self-learning ranking models and advanced analytics. This requires the user's clickstream.
- Optimization for media. Create recommendation and search apps optimized for media content.
- Natural language querying of healthcare data. Search FHIR resources without prior knowledge of any query language.
- Context-aware healthcare searches. Find search results with semantic
relevance that a structured FHIR search might miss.