Implementing Multi-Party Video Conference Escalation for Complex Case Resolution

How-To-Articles
1

Implementing Multi-Party Video Conference Escalation for Complex Case Resolution

What This Guide Covers

This guide details the architectural pattern for bridging a Genesys Cloud CX voice interaction with a WebRTC video session to include external subject matter experts (SMEs) and system integrations. The end result is a seamless handoff where a voice agent invites a video-capable SME to join the active call via the Genesys Cloud Agent Desktop, allowing for screen sharing, annotation, and real-time visual collaboration without dropping the customer connection.

Prerequisites, Roles & Licensing

Licensing

  • Genesys Cloud CX 2 or CX 3: Required for WebRTC video capabilities and advanced routing.
  • Video Add-on: Explicitly required for all agents and SMEs participating in the video leg. Voice-only participants do not require this license, but the session initiator must have it.
  • WEM Add-on (Optional): Required if you intend to record and analyze the video session for quality assurance.

Permissions & Roles

  • Telephony > Trunk > View/Edit: To configure the SIP trunk supporting the video codecs.
  • Routing > Queue > Edit: To configure the escalation queue with video capabilities.
  • Interaction > Interaction > Edit: To manage the interaction data model for multi-party merging.
  • User > User > Edit: To assign the Video license and enable the WebRTC softphone for SMEs.

External Dependencies

  • WebRTC-Supporting Browser: Chrome, Edge, or Firefox for SMEs.
  • SIP Trunk with Video Codecs: Your PSTN carrier or SIP trunk provider must support H.264 or VP8/VP9. If the customer is on a legacy landline, the video leg is internal only (Agent to SME).
  • Network Policy: Firewalls must allow UDP ports 3478 (STUN) and the media port range (typically 10000–20000) for WebRTC traffic.

The Implementation Deep-Dive

1. Configuring the Video Capable Queue and Routing Strategy

The foundation of this architecture is a dedicated queue that signals to the Genesys Cloud platform that this interaction requires video resources. Standard voice queues will strip video attributes during the routing process, causing the handoff to fail.

Step 1.1: Create the Escalation Queue
Navigate to Admin > Routing > Queues. Create a new queue named Escalation - Video SME.

  • Skills: Assign a specific skill, e.g., Video_Escalation.
  • Video Enabled: Toggle Video to ON. This is the critical configuration key. Without this, the queue cannot accept or route video-capable interactions.
  • Wrap-Up Codes: Define codes that allow for “Video Session Closed” to ensure accurate reporting in WEM.

Step 2.1: Configure the Routing Strategy
In the same queue settings, navigate to the Routing tab.

  • Overflow: Set a reasonable threshold. Video SMEs are scarce resources. If the queue depth exceeds 5, consider routing to a voice-only expert or a callback queue to prevent excessive wait times.
  • Longest Idle: Select this mode. Video sessions are resource-intensive. You want the SME who has been idle the longest to ensure fair distribution and reduce context-switching fatigue.

The Trap: Codec Mismatch on the Trunk
If your SIP trunk does not explicitly negotiate H.264, the Genesys Cloud platform will downgrade the interaction to voice-only when bridging the customer.

  • Misconfiguration: Leaving the trunk codec preferences at “Default” or “G.711 only”.
  • Downstream Effect: The SME joins the call, but the customer sees a static screen or receives a “Video Not Supported” error. The interaction remains active, but the visual collaboration fails.
  • Solution: Verify your Admin > Telephony > Trunks settings. Ensure H.264 is listed in the Video Codecs section and is prioritized. For WebRTC internal legs, VP8 is preferred for lower latency, but H.264 is the universal bridge for PSTN-to-Video scenarios if the carrier supports it.

2. Architecting the Multi-Party Flow in Genesys Cloud Architect

The core logic resides in the Architect flow. You must construct a flow that captures the voice interaction, escalates it to the video queue, and manages the state of the interaction until the SME joins.

Step 2.1: The Escalation Trigger
Start with a Queue block.

  • Queue: Select Escalation - Video SME.
  • Interaction Type: Set to Video. This forces the platform to attempt a video connection.
  • Timeout: Set to 300 seconds (5 minutes). Video SMEs are not always available. A long timeout will tie up the agent and the customer.

Step 2.2: Handling the SME Join
When the SME accepts the interaction, the platform creates a new leg. You must merge the legs.

  • Use a Transfer block or Conference block depending on your desired topology.
  • Recommended Pattern: Use the Add Participant action within the flow if you are using the API-driven approach, or rely on the Agent Desktop’s native “Add to Call” functionality if the agent initiates the SME pull.
  • For a fully automated flow:
    1. Queue (Voice) → Customer waits.
    2. Agent decides to escalate.
    3. Flow transitions to Set Variable: escalation_reason = "Complex Case".
    4. Queue (Video) → SME is routed.
    5. Conference: Merge Customer Leg, Agent Leg, and SME Leg.

Step 2.3: Data Passing via Interaction Attributes
You must pass context to the SME. The SME needs to know why they are joining.

  • Use the Set Interaction Data block before queuing.
  • Key: escalation_context.
  • Value: Concatenate the customer’s case number, the issue type, and the agent’s notes.
  • Payload Example:
    {
  "escalation_context": "Case #12345 - Payment Dispute. Agent notes: Customer claims double charge on invoice #998.",
  "priority": "High"
}
  • This data appears in the SME’s Agent Desktop sidebar, allowing them to prepare before answering.

The Trap: Dropping the Customer Leg During Handoff
A common error is using a “Transfer” action that completes the original interaction and starts a new one.

  • Misconfiguration: Using Transfer > Complete instead of Transfer > Warm or Conference.
  • Downstream Effect: The customer hears a disconnection tone. The original interaction is closed, and a new interaction ID is generated. This breaks the audit trail, loses the original IVR data, and resets the queue time metrics.
  • Solution: Always use a Conference block or a Warm Transfer that keeps the original interaction open. The SME joins the existing interaction as a new participant. This preserves the interactionId and ensures all legs are recorded under a single session.

3. Agent Desktop Configuration for SMEs

The SME experience is driven by the Genesys Cloud Agent Desktop. The SME must have the correct capabilities enabled to receive and initiate video.

Step 3.1: Enable WebRTC Softphone
Navigate to Admin > Users. Select the SME user.

  • Telephony: Ensure WebRTC is selected as the device type.
  • Video: Toggle Enable Video to ON.
  • Permissions: Assign the Telephony > Video > Use permission.

Step 3.2: Customizing the Desktop for Video
SMEs need quick access to screen sharing and annotation tools.

  • Settings > Preferences > Video: Set the default camera and microphone.
  • Custom Fields: Add the escalation_context field to the interaction sidebar. This ensures the SME sees the reason for the escalation immediately upon answering.

The Trap: Browser Compatibility and Permissions
SMEs often fail to join because their browser blocks camera/microphone access or lacks WebRTC support.

  • Misconfiguration: Using Internet Explorer or an outdated version of Safari.
  • Downstream Effect: The SME clicks “Answer,” but the video feed is black. The audio may work, but the collaboration fails.
  • Solution: Enforce a browser policy. Require Chrome 90+ or Edge 90+. Provide a standard operating procedure (SOP) for SMEs to allow camera/mic permissions in the browser settings. Test the setup using the Genesys Cloud > Settings > Test WebRTC tool before going live.

4. API-Driven Escalation (Advanced Integration)

If you are integrating with an external CRM (e.g., Salesforce) that triggers the escalation, you must use the Genesys Cloud APIs to modify the interaction state.

Step 4.1: Adding a Participant via API
When the CRM determines an SME is needed, it calls the Genesys Cloud API to add the SME to the existing interaction.

Endpoint: POST /api/v2/interactions/{interactionId}/participants

OAuth Scopes: interaction:write, routing:write

JSON Payload:

{
  "participantId": "sme_user_id_from_salesforce",
  "type": "video",
  "role": "agent",
  "state": "not-in-queue",
  "properties": {
    "escalation_reason": "CRM Triggered SME Request",
    "case_number": "CASE-998877"
  }
}

Step 4.2: Monitoring the Session State
You must poll the interaction status to ensure the SME has joined.

Endpoint: GET /api/v2/interactions/{interactionId}

Response Check: Look for a new participant object with state: "in-session" and type: "video".

The Trap: Race Conditions in Participant Addition
If the agent manually adds the SME while the CRM is also attempting to add the SME via API, you may get a conflict.

  • Misconfiguration: No locking mechanism or idempotency check in the integration code.
  • Downstream Effect: The API returns a 409 Conflict error. The SME is not added, and the escalation fails. The customer is left waiting.
  • Solution: Implement a “lock” flag in the CRM record. When the CRM triggers the API call, it sets escalation_in_progress = true. If the agent manually initiates the escalation, the CRM checks this flag and aborts the API call. Alternatively, use the Idempotency-Key header in your API requests to prevent duplicate participant additions.

Validation, Edge Cases & Troubleshooting

Edge Case 1: Customer on Legacy PSTN (No Video Support)

The Failure Condition: The SME joins the call, but the customer cannot see the video. The SME sees a black screen.

The Root Cause: The customer is connected via a traditional PSTN line that does not support H.264 or WebRTC. The Genesys Cloud platform cannot bridge video to a voice-only leg.

The Solution:

  1. Detect Capability: In Architect, use the Get Participant block to check the customer’s capabilities.
  2. Conditional Logic: If capabilities.video is false, do not route to the video queue.
  3. Fallback Path: Route to a voice-only SME queue. The SME can still use screen sharing internally with the agent, and the agent can describe the visual information to the customer.
  4. Communication: Use the Set Interaction Data block to flag customer_video_capable = false. Display this in the SME’s desktop so they understand why they cannot see the customer’s camera (if applicable).

Edge Case 2: Bandwidth Throttling and Packet Loss

The Failure Condition: The video feed is choppy, frozen, or has significant audio-video sync issues.

The Root Cause: The SME’s network connection is unstable, or the corporate firewall is throttling UDP traffic. WebRTC is sensitive to packet loss.

The Solution:

  1. Network Diagnostics: Use the Genesys Cloud > Settings > Network Diagnostics tool for the SME’s device. Check for jitter and packet loss.
  2. Codec Fallback: Configure the queue to allow codec fallback. If H.264 fails, fall back to VP8.
  3. Resolution Adjustment: In the Agent Desktop, allow the SME to manually lower the video resolution. High-resolution video consumes more bandwidth.
  4. Firewall Rules: Ensure that the UDP port range used for WebRTC media is not being rate-limited by the corporate firewall. Work with IT to whitelist the Genesys Cloud media servers.

Edge Case 3: SME Unavailability and Queue Overflow

The Failure Condition: No SMEs are available. The customer waits indefinitely in the video queue.

The Root Cause: The video queue has no overflow strategy, or the overflow threshold is too high.

The Solution:

  1. Overflow Configuration: In the queue settings, set an overflow threshold (e.g., 3 minutes).
  2. Overflow Action: Route to a voice-only SME queue or a callback queue.
  3. Notification: Send an email or SMS to the SME group if the queue depth exceeds 2. This alerts SMEs to potentially become available.
  4. Customer Communication: Use the Play Prompt block to inform the customer: “We are connecting you with a specialist who can use video. If no specialist is available, you will be connected with a voice specialist.”

Official References