Read AI launches its Large Meeting Models (LMMs) to expand the context window from a single meeting to thousands of meetings.
“In the past, meeting summaries were like articles in a magazine, interesting, but rarely connected to anything else. With the launch of Read’s LMMs, meetings become chapters within a book, where context plays a key role in what happens next,” said David Shim, Co-Founder and CEO of Read.
Today, we are sharing the availability of innovations made possible with Read’s LMMs:
LMMs applied to meetings Read attended on your behalf or were shared with you, to generate a personalized podcast that highlights the past 24 hours and prepares you for your upcoming meetings. Prepare for the day by listening to the Daily Read during your morning commute. Sample Daily Read Podcast.
With access to thousands of your meetings, as well as those meetings shared with you, Read can discover common questions based on topics. Read can then pull in answers from past and future meetings based on popularity and reaction from participants.
A single meeting can create a list of action items, but it isn’t able to track progress and delivery. Read’s LMMs identifies an action item, and creates an assistant that scans through thousands of meetings, determines action items, and updates the status of those action items in real-time.
Instead of a single meeting summary, Read’s LMMs learn from your meetings what topics are of interest to you, search across thousands of pages of meetings, and generate a personalized For You Page. This page provides a 360-degree view of the topic, team or workspace, incorporating content from your meetings as well as those shared across your team and company.
Read AI’s LMMs are not limited to meetings, and today we’re introducing the addition of calendars. By incorporating calendars into Read’s LLMs, Read is able to provide meeting participants transparency on the best time to schedule the meeting using past engagement and sentiment scores.
The ability to expand beyond a single meeting by breaking through token limits and context windows enables Read to introduce new meeting modalities.
“Large language models (LLMs) alone are limited by their context windows and limited steerability. The largest public models are limited to approximately 100,000 words, which translates to approximately 10 meetings, while Read AI’s LMMs are designed for contexts that are 100x larger,” said Robert Williams, Co-Founder and VP of Engineering at Read. “Read’s LMMs is the federation of meetings, incorporating context from prior and future meetings to enable contextualization that is leaps and bounds above the rest of the market.”
Read is the system of record for meetings having crossed over 100 million minutes of meetings measured to date along with Zoom selecting the company as an Essential App, Google incorporating Read as a launch partner for Meet Add-Ons, and partnership with Microsoft Azure. With the introduction of LMMs, Read is establishing itself as the system of record for Productivity AI.