Three questions worth answering properly.
Longer than a blog post and written to be the last thing you need to read on the subject. Each one links down to the specific pieces underneath it.
What an AI meeting assistant actually does, and the four questions that separate them.
The category went from novelty to default in about eighteen months, and the products in it are now so similar on a feature grid that the grid has stopped being useful. The differences that matter are architectural, and none of them appear on a comparison table.
6 sections · 8 linked articles
Transcription that never leaves the device is no longer a research project.
Of fourteen meeting products surveyed in September 2026, exactly one besides Ledgeur ships any on-device processing at all, and that one is partial. This is a guide to why it is rare, why it is now possible anyway, and how to tell a real claim from a careful sentence.
5 sections · 7 linked articles
Your agent is guessing about your work, and the answer was in a meeting.
The Model Context Protocol is how an AI tool reads a system it does not own. Search demand for it runs at roughly 14,800 a month and nobody in the meeting-notes category has written the version of this that matters: what it means when the system being read is everything your company has said out loud.
5 sections · 4 linked articles