Feature Overview
Webex AI Agent is an AI-powered solution designed to create, manage, and deploy automated AI Agents to fulfil customer service and support needs. It is designed to enhance customer interactions by automating routine tasks, delivering prompt responses, and increasing overall customer satisfaction.
Webex AI Agent provides automated assistance to customers before they interact with human agents. These agents support voice interactions with intonation, language understanding, and contextual awareness within conversations. AI Agents seamlessly and informatively handle digital channel interactions on all supported digital channels. Customers benefit from a personalized service, receiving assistance with questions, information retrieval, and minimizing wait times.
The AI Agent experience includes two agent types: Scripted AI Agent and Autonomous AI Agent. Each agent type supports a different set of capabilities depending on the interaction model and the level of automation required. Scripted AI Agent is intended for guided, predefined conversational flows, while Autonomous AI Agent supports more dynamic and flexible interactions, including advanced event handling and transfer scenarios.
-
Scripted AI Agents are rule-based agents that use a predefined set of rules, Natural Language Understanding (NLU) frameworks and classical machine learning models to generate responses, allowing the developer to define the response the AI Agent uses for specific customer queries.
The Scripted AI Agent supports the following capabilities:
-
EU AI Act Compliance
Scripted AI Agent supports compliance-related requirements aligned with the EU AI Act. This helps ensure that AI-driven interactions can meet applicable transparency, governance, and regulatory expectations where required.
-
A-law Codec Support
Scripted AI Agent supports the A-law codec for voice interactions. This enables compatibility with deployments or telephony environments that require A-law audio encoding.
-
-
An Autonomous AI Agent that can generate human-like, conversational responses to a customer query dynamically by using a Large Language Model (LLM).
The Autonomous AI Agent supports the following capabilities:
-
EU AI Act Compliance
Webex AI Agent now includes compliance-related enhancements that support AI development, data privacy, security, and safety requirements for customer interactions. For more information, see the AI development, data privacy, security, and safety section in the Webex AI Agent Studio Administration Guide.
-
A-law Codec Support
Autonomous AI Agent supports the A-law codec for supported voice integrations. This allows the agent to be used in environments where A-law codec support is required.
For more information, see the following guides:
The Feature Overview section in the Webex AI Agent chapter of Cisco Unified Contact Center Enterprise Features Guide, Release 15.0(1) and Cisco Packaged Contact Center Enterprise Features Guide, Release 15.0(1).
-
Dynamic Welcome Message
Autonomous AI Agent now supports dynamic welcome messages, allowing greetings to be personalized based on context such as the customer profile, entry point, or interaction scenario. For more information, see the Update Autonomous AI Agent Profile section in the Webex AI Agent Studio Administration Guide.
-
Custom Events: Custom Exit and Re-entry
Autonomous AI Agent now supports custom events for exit and re-entry flows. This enables more flexible conversation handling when a customer leaves a flow and later returns, or when a specific event-based path needs to be triggered. For more information, see the Configure Custom Transfer Action section in the Webex AI Agent Studio Administration Guide.
-
Silent and Announced Transfer
Autonomous AI Agent now supports silent transfer and announced transfer. Silent transfer moves an interaction to another destination without requiring additional customer action, while announced transfer provides for an announcement before completing the transfer. For more information about silent and announced transfers, see the Configure custom transfer action section in the Webex AI Agent Studio Administration Guide.
-
MCP Fulfillment Actions
Autonomous AI Agent now supports Model Context Protocol (MCP) fulfillment actions, enabling AI agents to connect directly to third-party tools and services during live conversations. For more information, see the Configure MCP client action section in the Webex AI Agent Studio Administration Guide.
-
Web URLs as Knowledge Sources
Autonomous AI Agent now supports web URLs as knowledge sources, allowing administrators to extract content from authorized public websites and use it for AI Agent responses. For more information, see the Create web URL knowledge source section in the Webex AI Agent Studio Administration Guide.
-
Webex AI Pro Europe Engine
Autonomous AI Agent now supports the Webex AI Pro-Europe 1.0 engine, providing a localized experience for Europe-based customers. For more information, see the Update autonomous AI agent profile section in the Webex AI Agent Studio Administration Guide.
-
Webex AI Agent offers the following feature specifically for voice channels:
-
AI Agent Call Transcription: Provides detailed transcripts of interactions between end customers and the Webex AI Agent, capturing every aspect of the conversation. This ensures precise documentation, so no information is missed or misinterpreted. The comprehensive transcripts allow businesses to analyze interactions more thoroughly, identify trends, recurring issues, and areas for improvement. They also help organizations better understand customer preferences and pain points, enabling more personalized and effective service.
Note |
CCE supports AI Agent capabilities for digital channels by providing agents with access to the customer interactions, rather than delivering real-time transcription of the conversation. This functionality is available across a variety of digital engagement channels, including WhatsApp, SMS, Facebook Messenger, Apple Messages for Business, and Live Chat. The feature automatically records and displays the entire interaction between customers and AI agents, allowing live agents to seamlessly review the interaction directly from their desktop interface. This comprehensive context equips agents with valuable insights, enabling them to provide more relevant and informed responses when they join the conversation. |
Note |
Virtual Agent Voice (VAV) supports both G.711 μ-law and G.711 A-law codecs. However, A-law codec is not supported for Google Dialogflow CX (DFCX) via Cloud. |
Feedback