Last updated 3 November 2026
What the AI does in Aveli
The assistant answers questions using the knowledge documents your team has approved, generates drafts and summaries, and helps produce follow-up reminders and reports. It runs on third-party large language models accessed through our AI gateway.
What it is grounded in
Answers are built from the documents your team has uploaded and approved for this workspace, plus the conversation you are in. If a fact is not in that material, the assistant is told to say so rather than invent it. You control which documents are approved, and removing a document stops it being used.
Limitations you should know about
AI answers can be wrong, outdated, or incomplete, and can sound confident while being wrong.
The assistant can only see what you have given it; it does not know your business unless you tell it.
Outputs are not regulated professional advice and should not be used for medical, legal, financial, or similar decisions without a qualified human review.
Automated follow-ups and reminders are suggestions generated from your own data and rules; you decide what is actually sent.
Your responsibilities
You choose what to publish to customers, you are responsible for the accuracy of anything you send, and you should tell your customers when they are interacting with an automated system where your market or law expects that. Keep a human in the loop for anything consequential.
What we do not claim
We do not claim that AI output is verified, that every answer has been reviewed by a person, or that the assistant is suitable for regulated decision-making on its own.