Configure Google Dialogflow ES to Improve Speech Recognition for CCE VA Implementation

Introduction

This document describes how to configure Google Dialogflow Essentials (ES) to enhance Cisco Contact Center Enterprise (CCE) Virtual Agent (VA) speech recognition.

Contributed by Ramiro Amaya and Maria Jose Mendez Vazquez, Cisco TAC Engineers, and Anjum Jeelani and Raghu Guvvala, Cisco Engineers.

Prerequisites

Requirements

Cisco recommends that you have knowledge of these topics:

• Cisco Unified Contact Center Enterprise (UCCE) Release 12.5
• Cisco Package Contact Center Enterprise (PCCE) Release 12.5
• Google Dialogflow ES

Components Used

The information in this document is based on these software versions:

• UCCE Release 12.5
• Customer Voice Portal (CVP) 12.5
• Virtualized Voice Browser (VVB) 12.5.1 ES12
• Google Dialogflow ES

The information in this document was created from the devices in a specific lab environment. All of the devices used in this document started with a cleared (default) configuration. If your network is live, ensure that you understand the potential impact of any command.

Background  Information

There are several scenarios in which a caller provides information to a virtual agent and the virtual agent misinterprets the information. For example, the virtual agent requests identification or member id, a credit card number, a phone number, and so on, and the virtual agent does not recognize or misinterprets the information provided by the caller. When you troubleshoot this scenario, you must consider a specific configuration on the Dialogflow ES recommended by Cisco and Google.

Configure

Before you configure Dialogflow ES in order to improve the speech recognition, ensure that you have installed the latest VVB Engineer Special (ES). Then, proceed to make changes to the Dialogflow ES configuration.

Step 1. Navigate to the Dialogflow ES console, then to Settings > Speech. Under Improve Speech Recognition Quality, ensure that Enable Enhanced Speech Models and Data Logging and Enable Auto Speech Adaptation are enabled.

Step 2. Navigate to the agent’s Intents panel, look for the intent that matches the caller's conversation. Under Action and parameters, mark the regexp entity as a required intent parameter, so that it can be collected slot-filling, as shown in the example. This allows auto speech adaptation to be strongly biased for sequence recognition so it does not try to recognize intent and sequence at the same time.

Verify

There is currently no specific information to verify for this configuration.

Troubleshoot

There is currently no specific troubleshoot information available for this configuration.

Related Information

• Google Speech Enhanced Model
• Enabling data logging and using enhanced model
• Technical Support & Documentation - Cisco Systems