Add a Smart Search on Your Website with Google AI

There is increasing interest in AI technology, and one practical application already available for websites is the Google AI Agent, which Enonic has integrated to enhance search functionality.

This system provides a more advanced, intelligent search experience by leveraging generative AI and customized content indexing.

See a video of our recent presentation on the matter (in Norwegian):

Google AI Agent and RAG Technology

One widely recognized generative AI model is GPT (Generative Pre-trained Transformer), which generates answers to user queries based on pre-trained data.

However, another effective method is RAG (Request Augmented Generation), which tailors AI responses using content specific to a database you provide. This approach ensures that the AI generates precise, relevant answers instead of general ones, drawing from your custom data sources.

The RAG process involves sending a user query to the AI, which then retrieves relevant information from a database—often a vector database.

A vector database organizes content by associating related information, enabling the AI to find the closest matches to a query. This is similar to Google Image Search, where similar images are displayed based on a user’s input.

While Enonic has not yet implemented vector search, efforts are underway. In the meantime, we have developed a solution that pulls and indexes content from websites, such as our developer portal, as a foundation for AI-generated answers.

Implementation in Practice

To implement this technology, the Enonic team indexed the developer portal by configuring a Google AI Search Agent. This involved selecting an appropriate AI model and integrating the functionality into the system.

Businesses can use an app developed by Enonic, available on the Enonic Market, or opt for custom adaptations to deploy this search feature on their websites.

Once the agent is configured, users can ask questions like "How to get started with Enonic?" The AI will respond using information from the indexed portal and display the database results it relied on. This transparency allows users to verify the sources of the AI’s answers.

See also: Behind the Scenes of Enonic's New AI »

Benefits and Customization

As mentioned, Enonic offers an app that integrates AI-powered search into websites, allowing end users to access intelligent search functionality. The app is highly customizable, enabling businesses to adjust visual elements such as icon sizes or text boxes. It can also support follow-up questions, creating a more interactive user experience.

The Google AI Search Agent is not limited to public websites; it can also be used in closed systems like intranets, provided the AI has access to the necessary data. While the initial indexing process may take time—up to a day—it continuously updates afterward. Once indexed, the system can deliver highly specific, AI-generated answers or replace traditional search functionalities.

Continuous Improvement

Enonic emphasizes the importance of maintaining high-quality databases to ensure accurate and helpful AI responses. Users are encouraged to provide feedback and suggest improvements, such as integrating more code examples from GitHub.

While the setup process is relatively straightforward, the effectiveness of the AI search heavily depends on the quality and relevance of the indexed content.