Versadial Call Recorder speech analytics 

The feature is available starting with VSLogger revision 4.8.6.0 (August 2018). The initial release supports analytics engines from  GOOGLE and  Yactraq.

Now that you have your calls recorded, listening to them individually can be time-consuming! Take the guessing game out of monitoring calls. Have them systematically reviewed using speech analytics. Speech analytics uses search algorithms that analyze thousands of minutes of audio searching for specific phrases to categorize and flag before calling a human to efficiently review.

HOW IT WORKS:

Versadial’s call recorder, VSLogger, now allows audio to be uploaded to speech analytics engines from multiple third-party providers.  (Additional third-party Speech Recognition services to be announced!).

Where Versadial comes in:

• We’re able to automatically upload your voice recordings to a speech analytics service of your choice
• Upload selected recordings to a speech analytics service of your choice on demand
• Display transcription within your Playback interface or automatically email to assigned users or supervisor
• Automate scheduled actions (EX: automatically flag a call that uses the word “discount,” then e-mail it to a supervisor)

Our summer Speech Analytics ability includes:

• Speech-to-text conversion
• Search-by-word or phrase from an audio recording
• Flag, bookmark, or e-mail calls containing YOUR custom criteria: (EX:  flag, bookmark, or e-mail calls that contain “credit card”
• Automatically erase personally identifiable information (credit card information format to maintain PCI compliance)

Coming soon:

• Voice recognition software: auto-identify the speaker

Q1. How do I get started?

A. All you need is two things.

1. Versadial’s Analytics Capability License (this allows us to enable VSLogger to work with your speech analytics solution on the back-end)
2. Speech Analytics Subscription of your choice (We’re currently integrated with Google and Yactraq, but pick any service you want, and contact us to confirm compatibility!) (Note: More than one Provider can be configured/supported at the same time)

Q2. I'm a current Versadial customer and was wondering if there's a way to determine what the analytics services would cost based on our current usage?

A.  Versadial doesn’t control analytics providers fees. Please review this page  for provider fees info  https://www.versadial.com/speech-analytics-for-your-recorded-calls/

You can control your analytics spending by using only manual (on demand) submissions of select recordings to Analytics processing.

To estimate spending for automatic submission of all recordings to Analytics processing:

• run a search for an average work day (24 hours, all channels).
• Estimate daily  total recording hours per day 
• multiply on number of working days per month.

e.g. 10 ch recorder, avg. 2 hour per day per channel,  20 work days  = 400 hours per month. Now you can estimate total monthly fee for individual providers,  e.g. Google 400 x 1.44 = $576

Mixed Audio Sample results screen

Separate TX/Rx streams Audio Sample results screen

---

ANALYTICS   CONFIGURATION  

Step 1:  Select the desired Analytics provider,   configure Provider API account/ subscription (e.g. Google Cloud).  Configure Versadial recorder to use provider API

SUPPORTED PROVIDERS	VSLogger Configuration
Google Cloud Visit provider site Speech to text conversion powered by machine learning and available for short or long-form audio Google Cloud Speech-to-Text enables developers to convert audio to text by applying powerful neural network models in an easy to use API.  The API recognizes 120 languages and variants, to support your global user base. You can enable voice command-and-control, transcribe audio from call centers, and more.  It can process real-time streaming or pre-recorded audio, using Google’s machine learning technology. Main Features Powered by machine learning. Recognizes over 120 languages. Inappropriate content filtering.  Automatic speech recognition How to open the account  Go to  https://console.cloud.google.com/home See configuration tips below, on this page.	WHERE:  Settings/ Voice Analytics  section PROVIDER: GOOGLE, Press [>] button, Press Edit in drop down panel Recorder Index: 1  (2 etc. if you have more than one recorder using the same Account) Google Cloud Storage Bucket:   enter value configured in Google account Default Language:    Speech Language (by default) Channel Language:   You can assign a different language to selected channels Google Cloud Private Key:   select JSON file provided by Google when you configured  the provider account Press  [Submit]  button to save
YACTRAQ Visit provider site  The cutting edge in speech analytics and audio mining, Yactraq empowers Call Centers to take full control of their business with machine learning driven insights extracted from any call recording. By utilizing our advanced technology to generate customized speech vocabularies that are configurable to a client’s specific business and needs, Yactraq provides a suite of call operations solutions that offer a high ROI, have big data applications, and can shine a comprehensive light on highly valuable dark data Main Features Agent performance evaluation. Call operations management. Compliance. Gartner 2016 cool vendor.  How to open the account  Contact provider for a new account.  Mention that you want to open an account for use with the Versadial recorder. Provider reps should provide information required for VSLogger configuration	WHERE:  Settings/ Voice Analytics  section PROVIDER: YACTRAQ, Press [>] button, Press Edit in drop down panel Encryption: Currently NOT used Cipher: Currently NOT used Key: Currently NOT used Recorder Index: 1  (2 etc. if you have more than one recorder using the same Account) CLIENTID:   enter client Id AWS S3 Bucket:  enter the configured bucket AWS S3 Region:  SELECT REGION AWS S3 Access Key Id:  ENTER ID AWS S3 Secret Access Key:  enter configured key Press  [Submit]  button to save

 

Step 2:  Estimate analytics requirements (hours of transcribed audio per month). EX: “500 hours per month”

a) Contact Sales@Versadial.com
b) Purchase Versadial’s Analytics Capability License

The new capability will be enabled with license reactivation (see.  VSLogger UI,   Settings/ License)

Step 3: Schedule Analytics to Run (open “Settings” tab in VSLogger)

In VSLoger user program, select  Settings/ Voice Analytics panel.  Right below individual analytics providers (GOOGLE and YACTRAQ), there is a panel displaying analytics tasks in progress.

To enable automatic  submissions of the recordings for analytics processing, schedule  periodic analytics task as follows:

• Press [Scedule Task]  button.   "Schedule task"  dialog will popup.
• Provider: Select service provider (Google, Yacktraq etc)
• Period:    Select active period   "Hourly (for the previous hour)"
• Starting from:  select  date and time (24-hour format, 00..23)
• Storage:  Select recording storage to be used as a source of the recordings analyzed.  Main Storage is recommended.
• Channels: select channels selected for the analytics processing ( as comma delimited list of channel ranges,  e.g. 1-4,7,25-28 )
• Duration >:  select minimum recording length, in minutes. recoding shorter than this value will not be submitted for processing
• Expire After:   schedule process to expire after the following period  ( Never, 3 months, 6 months, 1 year  )

Task Processing hints:

• Each task Task has the following stages/states.  0-Prepare, 1-Upload, 2-Get Result, 3-Postprocessing.
• Upon completion of the current stage, each task starts one or more tasks for the next stage and terminates processing with one of the following status (FINISHED or FAILED)
• Scheduling one task will result in the continuous sequence of  tasks PENDING and Completed tasks, which can be monitored and reviewed by recorder Administrator

From this point, the VSLogger call recorder will send your audio recordings to your analytics service. The analytics Service provider will bill you for your API usage directly.

---

---

How to open and configure Google Cloud API account 

Go to  https://console.cloud.google.com/home

Main Menu access left top corner Icon (three Horizontal Bars)


1. Create a Google Cloud Platform project 

• Main Menu
• Select "IAM & admin" / Manage resources .   
• Press   [+CREATE PROJECT]
• Enter  Project name e.g.  Voice Analytics Project

2. Create a Service Account for the project

• Main Menu
• Select  project,
• Select "IAM & admin" / Manage resources .   
• Select Service accounts  
• Press [+CREATE SERVICE ACCOUNT]
• Enter Service Account Name e.g. Voice Analytics service 
• Service account ID will be generated  based on service name e.g. voice-analytics-service
• Select Role -  Owner
• Press [Save] 

3. Create Credentials (Service account key) download service account's public/private key  as a JSON file

• Main Menu
• Select  project,
• Select "IAM & admin" / Manage resources .   
• Select Service accounts. Click/select your service account. 
• Press [Edit]. [+ CREATE KEY] button will show
• Press [+ CREATE KEY] button,  select JSON, Press CREATE
• Key file will be created and automatically downloaded.
• Save the file and note file location. You will need this file to configure Versadial recorder

4. Enable Cloud Speech and Cloud Storage APIs

• Main Menu
• Select the project,
• Select "APIs and Services" / Library
• Find and Select "Cloud Speech API". Press Enable  
• Press [<- API LIbrary]  to return
• Find and Select "Google Cloud Storage" API. Press Enable  

5. Create a bucket  

• Main Menu
• Select the project,
• Select "Storage" / Browser
• "Sign up for an account"
•  Press [+ CREATE BUCKET ] . Remember the bucket name. You will need the bucket name to configure Versadial recorder.

From this point, you are ready to configure Versadial recorder to send recordings to your Google Cloud Analytics service.
Google will bill/charge you directly for the API usage.