NATURAL LANGUAGE UNDERSTANDING
A technology that lets machines understand human voice or written text, available on phones and the web. Ask us how IM Virtual Agents powered by NLU can improve customer experience.
Web video Agent
WHO IS THE END USER?
Italian citizens looking for information on the INPS web site.
WHAT IS THE PROBLEM?
Give to users access to INPS data using different channels, integrating many input channels (web, voice, web services) with different kinds of outputs (send email, send a video stream, send an MMS, etc…)
HOW WAS IT SOLVED?
Using IMSL Multimodal framework it is possible to collect the input from many possible channels:
- Voice recording from web
- Traditional call from PSTN
- Voip call
As each channel is identified by an unique identifier, it is possible handle concurrent input from different channels.
The demo extracts text from the audio data using a Language Model built using the customer data, then web can process the text with a semantic extraction process.
Finally, the result is sent to the user, using different output methods related to the channel originally used by the user.
The aim of this call is to give a vision of how our solution can extract keywords and concepts from a live call and push information to a call-center agent or directly to the caller
Give to users access to INPS data using different channels, integrating many input channels (web, voice, web services)