AI-Enabled Company Oracle
Developing a 'Company Oracle' chatbot for a financial technology firm, aimed at revolutionising information retrieval and decision-making processes within the company.
The Opportunity
Our client, a financial technology company, recognised the need to enhance their staff's ability to access company documentation and information. The goal was to create a 'company oracle'—a sophisticated chatbot powered by Large Language Models (LLMs) and integrated with the company's documentation through retrieval-augmented generation. This would enable the staff to quickly obtain information through conversational interactions with the chatbot, bypassing the slower process of manual searching.
What we did
✔︎ Developed a state-of-the-art company oracle chatbot with Retrieval Augmented Generation (RAG), designed to interface seamlessly with company resources. ✔︎ Implemented a user-friendly interface using Streamlit, facilitating effortless interaction for users. ✔︎ Supplied the client with the complete codebase and comprehensive deployment instructions to ensure a smooth transition to the new system.
The Results
The 'Company Oracle' chatbot significantly streamlines the research and decision-making processes by consolidating information from all facets of the business into a single, accessible knowledge base. It responds swiftly and accurately to inquiries, and includes functionalities for document management within the oracle's knowledge base. Users can also contribute feedback directly through the web application.
90%
retrieval accuracy in delivering prefect Oracle answers first time.
How we did it
Our approach began with an exhaustive data collection phase, amassing both internal and external resources pertinent to our client's business operations. This data was organised within a vector database, enabling the LLM-powered chatbot to perform retrieval-augmented generation across the documents. To determine the most effective agent, we conducted a series of tests and refined multiple iterations, ultimately selecting the agent that demonstrated superior performance. The chatbot has been integrated into a web application, where user interactions and feedback are meticulously logged. This data collection facilitates an ongoing improvement cycle for the chatbot's responses. All system components have been securely hosted on the Azure platform, ensuring robust data protection and system reliability.
AI-Enabled Company Oracle
Start a conversation
Take the first step by speaking with one of our data experts today.