This demonstration looks at using Intelligent Automation to feed a knowledge base with up-to-date content & data for use by Generative AI models. This solution addresses several concerns that are holding back adoption of Generative AI in the enterprise, including understanding the source of data used to generate a response, ensuring only relevant content is used in the creation of a ChatBot reply, allowing fact checking of AI generated responses, avoiding AI "hallucinations" & giving the AI model detailed understanding & context of the business domain in question.
The solution makes use of various techniques to augment the base training of the Large Language Model (LLM), including automated sourcing of content, Intelligent Document Processing, providing an AI Knowledge Base, Retrieval Augmented Generation (RAG), & Hybrid (Vector) Search.
The demonstration has been implemented using: Kofax TotalAgility, Kofax RPA, Azure AI Search, OpenAI GPT4 & a custom ChatBot UI.
For part 1 of this demonstration, see: Zero Code Research Automation with TotalAgility and OpenAI
