A Low-Code Framework for Conversational AI Assistants Leveraging Retrieval-Augmented Generation and LLMs

With the recent advancement in the field of Artificial Intelligence in the development of the large
language model which paves way for the rapid improvement of the conversational AI allowing the
smart and context aware interaction between the humans and the computer in a variety of fields. These
models demonstrates the ability of understanding the natural language and generates the human like
response to the user queries. Many business applications has the potential to make use of these
conversational AI for effective usage. But Integrating this into their applications is still difficult and
requires a high level of technical expertise to incorporate these AI into their application, since it requires
large amount of programming work, AI knowledge and system modification. To overcome this
limitation, this paper presents a low-code framework approach of creating a domain specific
conversational AI assistant with the help of LLMs and combining with the retrieval augmented
generation (RAG). By using the low code platforms like Oracle APEX, the developers or the business
users can easily create and integrate the smart assistants into their business applications with less
technical expertise. By combining the generative AI along with the retrieval-based methods, the
assistants can deliver the accurate response by referring the business knowledge sources like documents,
database etc.