The quality of the answers generated by Aptus.Chat is primarily linked to two factors. The first one is the machine-readable format of the documents available on our platform: by transforming the document in a format readable by the AI we are capable of providing the LLM systems with non-textual elements, in addition to the text content, which enhance the context available to generate an answer. The second main factor is the RAG (Retrieval-Augmented Generation) process, according to which our AI models identify the text law part(s) more pertinent to the question asked by the user to be used by the LLM system to generate the answer: in this way, the answer generated by Aptus.Chat is proven to be a re-elaboration of official regulatory content. Furthermore, it should be noted that the quality of the answers is strictly related to the quality of the questions formulated, in terms of input provided, user’s specific context and how questions are written.
