Chat Aible now has three chat flows:
- Analytics - where the LLM describes and queries the Aible Information Model to discover insights in your data without sending the raw data to the LLM and with hallucination protection.
- Direct to the LLM - where you can query your choice of LLM directly without additional prompt augmentation
- and now Document Reference and Summarisation
With Document Reference and Summarisation we securely upload PDF documents into your cloud account and build a vector database (also in your cloud account) representing the content in a format that supports natural language search such that similar or closely related concepts and ideas in the text can be easily discovered.
Users can search the documents via ChatAible to find content related to a query. ChatAible sends only the small subset of content found to the LLM for summarisation in much the same manner as the Aible Information Model is used for analytics: ChatAible analyses and summarises your data (or text) and sends only an abstract specific to the question to the LLM so summarise, never sending the raw data in order to maintain data privacy.
To get started with Document Reference and Summarisation in ChatAible speak to the Aible Customer Success team.