Launching an Agentic RAG Chatbot with MongoDB and Dataworkz

What Is an Agent?

An agent is an artificial computational entity with an awareness of its environment and associated data within the context. These agents can be used to interact in the case of ecommerce or perform a portion of complex tasks as needed.

Retrieval-Augmented Generation (RAG)

Retrieval-augmented generation is an approach used to augment large language models (LLMs) with proprietary data so that they can generate more accurate and context-aware responses while also reducing hallucinations.

Figure 1. Conventional RAG

Agentic RAG

Agentic RAG takes this a step further by introducing an AI agent-based implementation of RAG. In this model, different tools and functions can be accessed by the agent, enabling it to go beyond information retrieval and generation—it allows it to plan. Agents can determine if they need to retrieve specific information or not, decide which tool to use for the retrieval, and formulate queries. These capabilities are crucial as it enables the agent to pull information from multiple data sources, handling complex queries that require more than one source to formulate the response.

You can start to picture the enormous potential it has across industries to enhance customer interactions and streamline processes.

Figure 2. Agentic RAG

For more details see Dataworkz Agentic RAG.

Figure 3. Conventional vs Agentic RAG comparison

Impact on the Retail Industry

Generative AI is revolutionizing the retail industry. In 2024, one quarter of consumers used AI in their shopping experience, representing a 4% increase over the previous year. With such expansion, the potential for innovation remains vast.

Some generative AI use cases across retail include:

• Building customer-support chatbots: Imagine a real-time, GenAI-powered support chatbot that is able to assist customers at any point in time and is context aware of the business’s policies as well as the user’s history and preferences.

• Personalizing product recommendations: Providing the customer with recommendations based on their specific likes, needs, and past orders will make the shopping experience more targeted to the shopper making it more enjoyable and likely to end with a purchase.

• Creating dynamic marketing content: Generating fresh, relevant content for the customer by creating dynamic promotions, emails or messages specific for the customer. This leads to higher engagement, increased sales, and better customer retention.

All of the examples above show how, in everyday life, MongoDB and Dataworkz are already supporting retailers to build impactful customer experiences.

Introducing Dataworkz

Dataworkz is an enterprise-grade RAG-as-a-service platform that transforms how organizations build and deploy AI applications. At its core is an advanced, agent-based architecture combined with graph-optimized retrieval, enabling large enterprises to launch sophisticated RAG applications in hours, not months.

The platform eliminates the need for specialized AI teams through an intuitive no-code builder that automatically implements best practices in RAG development. Unlike traditional approaches that lock you into early architectural decisions, Dataworkz enables rapid experimentation—you can test different retrieval strategies, prompt variations, and model combinations in a controlled environment before committing to production.

From data ingestion to production deployment, Dataworkz handles the entire RAG lifecycle with enterprise-ready features for scaling, debugging, and operations.

Key differentiators:

• Graph-optimized knowledge retrieval for complex relationships.

• Agent-based architecture for sophisticated reasoning.

• No-code builder with built-in best practices.

• Full lifecycle support from experimentation to production.

• BYO (bring your own) flexibility—LLM, embedding model, vector database.

• Enterprise-grade security and scalability.

Trusted by Fortune 500 companies, Dataworkz delivers production-ready RAG applications without the traditional overhead of building and maintaining complex AI infrastructure.

Leveraging MongoDB with Dataworkz

​​By leveraging the strengths of both MongoDB and Dataworkz, organizations can create powerful, scalable, and accurate generative AI applications while minimizing development complexity and time-to-market.

Dataworkz streamlines the process of creating RAG pipelines by providing a point-and-click experience to extract unstructured data in any format, configure a chunking strategy, and create vector embeddings. Dataworkz RAG builder also allows developers to choose different retrieval mechanisms—lexical, semantic, or graph, with different thresholds to build the context for answering a user question.

The integration also enables real-time data processing and analytics, which is essential for generative AI applications that require up-to-date information. This capability ensures that AI models have access to the most current data, improving the accuracy and relevance of generated responses.

Agentic RAG Architecture

To build something transformational, conventional RAG systems, which primarily utilize vector search techniques, must be integrated with up-to-date information from operational databases.

MongoDB Atlas Vector Search provides built-in support for vector embeddings, which eliminates the need for a separate vector database, simplifying the architecture and reducing complexity.

That’s where agentic RAG shines. With Dataworkz agents, retailers can provide governed access to MongoDB collections, which are configured as tools. In many cases, customers have an API layer that abstracts the underlying collections—Dataworkz can be integrated with REST API or GraphQL.

Additionally, any RAG pipeline configured in Dataworkz can be made available as a tool to an agent. This gives agents the ability to understand unstructured data in a SharePoint site, a confluence wiki page, or markdown depending on the user’s question.

Existing retailers with applications that leverage MongoDB as their data platform will benefit from Dataworkz’s close integration with MongoDB and leap-frog their AI adoption to incorporate agentic RAG into their solutions.

Dataworkz agents can access multiple data sources and use reasoning LLMs to decide which tool to use to answer a user’s question. An agent can dynamically switch between different MongoDB collections or even multiple databases to bring in structured data such as shipping status, customer profiles and preferences, and order histories. Additionally, third party solutions like ERP (Epicor), CRM (Salesforce), and more can be integrated through APIs exposed via the providers. Internal services exposed as GraphQL is another popular type of integration. Together, these tools enable the agent to understand the user’s questions in their context and provide very personalized responses relevant to their experience and their data.

The way a Dataworkz Agent works is:

1. Dataworkz agent framework partitions use cases into scenarios, such as being able to answer questions about store policies, provide recommendations, or search for an order. Scenarios have a curated set of tools relevant to the scenario. When a user interacts with the agent, a classifier selects the appropriate scenario.

2. The agent then employs a reasoning LLM to come up with a plan to answer the question. Given the user’s question, the conversation context, memory, and the tools (such as access to a MongoDB collection) in the scenario, the LLM comes up with a sequence of steps to retrieve information to answer the question. This process is iterative; after each iteration, it’s a good idea to conclude by determining if enough context has been retrieved via the tools or if more iterations, planning, and execution are required to build more context.

3. Finally, the context constructed in step 2 is used by the agent to generate the necessary response or perform any action, or it might decide that it actually needs to ask the user a clarifying question.

Dataworkz brings the same ease of use to agents as it has to building RAG applications. Instead of requiring coding with libraries or frameworks, Agents are created by simply describing their intended functionality. They can be built declaratively, tested using built-in tools, and seamlessly deployed via an API when ready. This approach makes agents far more accessible and significantly more cost-effective to develop and deploy.

Figure 4. Dataworkz & MongoDB Agentic RAG architecture

MongoDB Developer Data Platform

• MongoDB Atlas

• Atlas Vector Search

Partner Technologies

• Dataworkz

• NextJS