Billing for licensed features

Gatekeeper has various licensed features available, which are licensed for a defined level of usage. This section describes how the usage of these features are tracked for billing.

Track usage through transactions

Most billing in Gatekeeper works by tracking transactions for audio authentication and fraud detection. A transaction is the action of processing audio or transcribed audio using a licensed feature. Customers have a fixed number of transactions per year, and using additional transactions may result in an extra charge per transaction.

For questions about your specific transaction availability, talk to your billing partner.

Sessions and consuming transactions

A channel is the source of audio for a session. The following are the two types of sessions:

• Online sessions are interactive sessions with the end customer (person). They can have multiple channels.
• Offline sessions are the activities of an application, system administrator or fraud analyst (when the person is no longer connected). They have a single channel consisting of entirely of audio from past calls.

A transaction occurs upon the first audio processing and the first transcription processing on each channel. If the first processing occurs in the online session, there’s no subsequent transaction during offline processing.

Processing that consume transactions

Each of the following are considered one transaction:

• Audio & Fraud Transaction (also called “Audio Transaction”): Audio processing, including one or more of:
  • All audio segments in the session/call, independent of the number of audio segments
  • Active authentication, passive authentication, or fraud detection
  • Deviceprint
  • Synthetic speech detection
  • playback detection
• Voice to Text Security Transaction: Processing audio for voice to text security, which analyses transcribed text
  • ConversationPrint, a behavioral biometric that recognizes patterns in conversation such as vocabulary, grammar, and sentence structure using transcribed text
  • Audio transcription

Example: One channel, online and offline

This example describes processing for one channel with online and offline sessions as follows:

1. Glinda connects to the call center for help with her new phone.
2. While on the call, Glinda is authenticated through voice biometrics.
3. After the call, the audio recordings and voiceprint are analyzed against a watchlist of known fraudsters, and transcribed.

Glinda’s call results in two transactions as follows:

• Audio Transaction for the audio processing to authenticate Glinda during the online session.
• Voice to Text Security Transaction for the offline transcription.

The post-call fraud detection has no fee because the prior Audio Transaction occurred during the live session authentication.

Example: One channel, offline

This example describes processing for a one channel offline session as follows:

1. Yulian connects to the sales center to buy insurance.
2. He speaks to an agent, but the agent does not use voice biometrics to authenticate.
3. After the call, Gatekeeper processes the recorded audio for fraud detection.

Yulian’s call results in one transaction for offline session audio processing. This occurs because there was no prior audio transaction during the live, online session.

Example: Two channels, online

This example describes processing for two channels and an online session as follows:

1. Xiang connects through the IVR channel to a tax center to ask a question about her tax bill.
2. The IVR uses passive voice biometrics to authenticate Xiang.
3. Later, the call transfers to an agent (Call Center channel).
4. The agent uses passive voice biometrics and ConversationPrint to authenticate Xiang.

Xiang’s call results in three transactions as follows:

• Audio Transaction for audio processing over the IVR channel
• Audio Transaction for audio processing during the Call Center Channel
• Voice to Text Security Transaction for ConversationPrint (which requires transcription) in the Call Center Channel.