In today’s rapidly evolving digital landscape, providing real-time, accurate, and contextually relevant customer interactions is essential but challenging. Traditional AI models are often limited by pre-training cutoffs, leaving a gap in timely knowledge delivery. Google’s Vertex AI Agent Builder, combined with Sitecore, offers a powerful solution to this challenge. Leveraging advanced machine learning models and Google Search Grounding, Vertex AI enables businesses to build dynamic, interactive AI applications that deliver current, context-driven information to enhance user engagement and satisfaction.
This blog dives into the solution’s framework and steps for building dynamic chat and search experiences with Sitecore and Vertex AI Agent Builder.
SOLUTION OVERVIEW:
The solution architecture employs a Retrieval-Augmented Generation (RAG) approach that gathers information from Sitecore and other sources, organizes it within a Data Store, and delivers conversational responses through an AI-driven interface. The primary components are Vertex AI Agent Builder, Sitecore’s SPE/Sitecore Connect modules, and Google Search Grounding, working together to create intelligent, knowledge-based, world-aware chat or search applications. This setup elevates customer engagement by providing timely, relevant responses to user queries.
Here’s a high-level diagram outlining the solution:
There’s no need to build the Chat/Search application and its underlying components from scratch. Vertex AI Agent Builder simplifies the setup, enabling quick deployment of the entire infrastructure in just a few clicks. The PowerShell module and the Connector SDK further streamline content formatting and integration with Vertex AI Data Store, which powers the chat and search capabilities.
SETUP INSTRUCTIONS:
Here are the high-level steps involved in setting up this module. For detailed instructions, please refer to the Github Repository.
- Set up a Google Cloud project and enable Vertex AI and Dialogflow APIs.
- Create a chat or search app in Agent Builder, linking an existing or new Data Store
- Configure the goals, instructions and tools
- Integrate Sitecore with Data Store to index website content
- For XP, deploy/customize the PowerShell module
- For XM Cloud, use the Connector SDK and Recipe with Sitecore Connect
- To enable Google Search Grounding, create a custom Google Cloud function to fetch, summarize, and return responses with data from Data Store and Gemini AI similar to this sample function app.
- Publish Sitecore content to populate the Data Store, validate agent responses, and embed the agent to your website using the Agent Builder’s publish option.
Source Code for this module is available on Github. This module is built for the Sitecore community and doesn’t require any license. This module doesn’t collect any of your information. Please check out and let me know if you have any feedback/issues/feature requests here. Thank you for using this module!