Addon Required
The Speech Analytics feature for Luware Recording requires an additional license addon for your environment. Contact solutions@luware.com or your customer success specialist for more information.
Overview
Speech Analytics for Luware Recording leverages cutting-edge machine learning, natural language processing (NLP), and artificial intelligence (AI) technologies to extract valuable insights from recorded conversations. By transforming raw audio data into actionable intelligence, it empowers organizations to enhance customer service, optimize operations, and bolster active surveillance efforts.
Powered by Intelligent Voice, Luware Recording Speech Analytics is a trusted solution for industries requiring the highest standards of security and compliance, including financial institutions and government agencies.
How does it work?
Speech Analytics powered by Intelligent Voice is natively integrated to Luware Recording with APIs providing a secure and effective method of speech analytics without data being sent to any third parties. Luware hosts the Intelligent Voice software within the perimeters of the Luware Recording Azure environment, following the same data privacy and security mandates as the Luware Recording platform.
When a conversation is captured by Luware Recording for a Speech Analytics enabled user, the media file is sent via an API to the Intelligent Voice software. Based on the license and configuration of the captured user, the Speech Analytics system will process the media file and generate an accompanying JSON file containing a transcript, sentiment analysis, topic analysis, translation and summary of the conversation.
The media file is then deleted from the Speech Analytics software and the generated JSON file is returned to Web Application for archive and storage. To enable text based search features a copy of the generated content is stored in the encrypted database. The generated JSON is signed and encrypted and uploaded to the customers storage account. The process usually completes within 30 minutes of the conversation being captured.
Supported Communication Platforms
Luware Recording Speech Analytics works with any audio files actively captured by the Luware Recording environment, independently of the communication platform. Only active licensed recorded users conversations captured by Luware Recording can be processed, for historical imports contract solutions@luware.com.
Key Features
Luware Recording Speech Analytics currently delivers the below features.
| Feature | Description |
|---|---|
| Speech Transcription | Converts audio files into transcribed text using Intelligent Voice language models. Also known as Automatic Speech Recognition (ASR). |
| Diarization | Identifies the number of speakers in each recording and segments the transcript to show which user is speaking. |
| Key Topics | Automatically identifies key topics discussed during a conversation and provides a link to that part of the audio file. Also known as JumpTo ™ Topics |
| Generative Summary | Automatically generates a summary of the transcript using artificial intelligence. |
| Automatic Language Detection | Automatically identifies the languages being spoken in each recording and can either transcribe in a dominant language or in multiple languages. |
| Multi-Language per user | Ability to assign multiple languages per user and identify the language being spoken for transcription. |
| Domain Language Training & Model Building | Language models can be adapted and tuned to understand customers specific internal language such as acronyms, lingo, or slang. |
Available Languages
Intelligent Voice adds additional languages based on customer requests, currently the below language models available:
Region |
Base Language |
Dialect |
|---|---|---|
| Global | English | International English |
| Europe | English | British English |
| Danish | Danish | |
| Dutch | Dutch | |
| French | French | |
| Finnish | Finnish | |
| Flemish | Flemish | |
| German | German | |
| German | Swiss German (Switzerland) | |
| Greek | Greek | |
| Italian | Italian | |
| Norwegian | Norwegian | |
| Polish | Polish | |
| Portuguese | Portuguese | |
| Russian | Russian | |
| Spanish | Spanish | |
| Swedish | Swedish | |
| Turkish | Turkish | |
| Americas | English | American English |
| French | Canadian French | |
| Portuguese | Brazilian Portuguese | |
| Spanish | Mexican Spanish | |
| Africa | English | South African English |
| Arabic | Egyptian Arabic | |
| Asia Pacific | English | Australian English |
| Cantonese | Cantonese | |
| Hindi | Hindi | |
| Indonesian | Indonesian | |
| Japanese | Japanese | |
| Korean | Korean | |
| Mandarin | Mandarin | |
| Tagalog | Tagalog | |
| Thai | Thai |
Differentiators
The Intelligent Voice software is built and trained for the enterprise market, with a special focus on recordings captured from unified communications in scenarios such as contact centre calls and financial trading. This combined with the deployment model at Luware, provides some key differentiators from using Public AI providers:
- Accuracy: The language models have been trained for the enterprise market with an additional focus on financial communication. This provides better accuracy when handling audio recordings, especially of low quality, which are common for enterprise use cases, such as contact centre calls, trading and unified communications.
-
Data Privacy: The language model is hosted by Luware in our Azure subscription, ensuring the below data privacy concerns are met:
- Models are not trained on customers data.
- Data is not available to any other customers.
- Data is not processed by any third parties.
- Data is not stored in the Intelligent Voice system.
- Data is not processed for content filtering.
- Data is not processed for abuse detection.
- No consumption billing: Luware maintains the infrastructure hosting the large language models in Azure, enabling us to offer the services as a simple per user per month price.
Enable Speech Analytics
Luware recommends the following sequence to enable Speech Analytics:
- Acquire the required license from Luware.
- Provide the languages to apply to each Azure Security Group.
- Luware will configure Speech Analytics.
Prerequisite Setup and Configuration is required
Luware Recording Prerequisites and setup must be completed before enabling Speech Analytics.
1. Acquire the License from Luware
Contact your Customer Success Specialist or solutions@luware.com to request the required Speech Analytics license. After the sales process is complete, the Luware customer success specialist will assign the license in the Luware Recording environment.
2. Provide the languages to apply to each Azure Security Group
Login to your Luware Recording environment and provide the Recorded User group name that should be enabled for Speech Analytics.
- Login to the Luware Recording Web Portal with a user with the Administrator role.
- Click Users > Groups.
- Copy the name of the Recorded User group.
- Fill in the table below with the Group names and language or languages that must be enabled for Speech Analytics.
Group Name |
Language/Dialect |
|---|---|
| LuwareRecording-RecordedUsers | English; Swiss German; |
- Provide this to your Luware customer success specialist.
4. Luware will configure Speech Analytics
Luware will now perform a series of changes, including updating the Active Directory Synchronisations for new users and bulk update existing users to enable Speech Analytics. On completion of these tasks, when playing back conversations, the Speech Analytics output will be available within the Web Portal.
Speech Analytics is enabled
After all the steps in this article have been completed, Speech Analytics is now enabled.