LumenVox speech products (NLU & diarization)
NLU
Natural Language Understanding (NLU) assists in understanding a speaker’s intent by using natural language processing for supplied text. This contains a number of subproducts.
Call Summarization
Instantly generate accurate, digestible call summaries to save time and improve record-keeping. The transcribed call text can be used to generate any of the following three product features available by calling the NLU service:
Call Summary
The user can set the configurations to summarize by number of words or bullet points required.
Topic Detection
Identify and categorize the main subjects discussed, from billing questions to technical support and more.
Outcome Detection
Clearly understand and record the result of each call, whether it's a sale closed, a task scheduled, or an unresolved issue.
Sentiment Analysis
This product utilizes the transcribed text to determine the sentiment of a sentence or entire call. The product is available for batch processing and produces one of the following responses:
- Positive
- Mostly Positive
- Mixed
- Neutral
- Mostly Negative
- Negative
- None
Language Detection
Language detection can be either text or voice based. Should you wish to use the text-based option then you simply submit text to the NLU service.
To use the voice-based detection, you need to call the language identification service
Language Translation
Language translation is consumed by submitting text to the NLU service. The input and output languages need to be specified but the input language can be automatically detected if the flag is enabled.
Speaker Diarization
Identify and separate speakers automatically in your audio recordings with advanced Speaker Diarization technology. This powerful tool leverages state-of-the-art speech processing algorithms to detect “who spoke when,” making it ideal for transcriptions, meeting analytics, podcast and media production, call centers, academic research, and more.
Speaker Diarization takes mono audio with multiple speakers and provides feedback on which speaker spoke when.