AMI research is developing technologies for processing the words spoken by participants during meetings and adding value. Unlike speech to text algorithms in use today, which require training, the speech to text systems for use in meetings will need to be relatively reliable without training or constraints as to vocabulary (Large Conversational Vocabulary Speech Recognition).

The audio/speech processing systems can also be used to:

• eliminate background noises in a meeting,
• remove silence (to accelerate the experience of the meeting), 
• identify the speaker,
• identify the language the speaker is using,
• spot key words,
• compress speech, and
• perform dynamic summarization.