Introduction

Dataworkz Agents are LLM-based multi-step software systems that leverage RAG and the ability to interact with the external world to enable use-cases far beyond what would be possible with conventional RAG. By incorporating decision making through powerful reasoning LLMs and a feature-rich and robust execution framework, Agents can solve complex problems, personalize results, interact with users with a degree of sophistication that enables myriad new use-cases.

Your business requirements and your imagination may be the limit to what you can achieve with Agents. Some use-cases that can be powered by Agents include -

Enhanced Information Retrieval (Advanced Agent driven RAG)

Dataworkz Agents can be deployed for internal employees going beyond conventional RAG to make the responses relevant to them or to their work. For instance, customer support agents can retrieve information about troubleshooting a customer’s issue that is aware of the customer’s specific product and version from the relevant system with the knowledge base of known issues. A conventional RAG system would be unable to retrieve the customer’s information and would provide more general information, requiring the agent to spend time deciding whether the answers are relevant or not.

Customer chatbot

Agents can act as a first contact chatbot on websites answering questions about their orders as well as applying policies and rules to their specific situation. The resultant answers from the Agent can be sophisticated and very personalized to the specific customer. This level of sophistication is not possible with conventional RAG.

This and several other use-cases can be achieved via Dataworkz Agents with minimal technical and AI knowledge.

In this document, we explain the concepts involved in Dataworkz Agents, dig deeper into some of the configuration aspects to understand them better and provide details on how you can create your own Agents.