How to Build a Custom Queue Fallback Cascade with Bullseye Routing

How to Build a Custom Queue Fallback Cascade with Bullseye Routing

Executive Summary & Architectural Context

In contact centers subject to strict SLAs, abandoning a caller in a hyper-specialized queue (like “Spanish IT Support for Product X”) for 45 minutes because the sole agent is at lunch is an unacceptable failure of routing architecture. While simple “If wait > 5 mins, transfer to General” logic exists, it creates terrible customer experiences, as the caller abruptly hears a ringing tone and is dumped into a massive generic pool, losing all context of their specific issue.

The engineering solution in Genesys Cloud is Bullseye Routing. Instead of transferring the caller physically between different queues (which restarts their wait timer and ruins queue analytics), Bullseye Routing expands the target audience of agents within the exact same queue over time.

This masterclass details how to architect a precise fallback cascade using Skills and Bullseye Routing rings, ensuring that a caller waits for the absolute best agent first, but gracefully expands to cross-trained agents without ever losing their place in line.

Prerequisites, Roles & Licensing

  • Licensing: Genesys Cloud CX 1, 2, or 3.
  • Roles & Permissions:
    • Routing > Queue > Edit
    • Routing > Skill > Add
    • Architect > Flow > Edit (to attach the initial skills)
  • Platform Dependencies:
    • A comprehensive taxonomy of ACD Skills (e.g., Product_X_Expert, Product_X_Generalist, Support_Overflow).

The Implementation Deep-Dive

1. Architectural Skill Definition

Bullseye routing relies entirely on the presence (and gradual removal) of ACD Skills. You must define a hierarchy.

  1. Ring 1 (The Experts): Require Language: Spanish and Skill: Product_X_Expert.
  2. Ring 2 (The Cross-Trained): Require Language: Spanish and Skill: Product_X_Generalist.
  3. Ring 3 (The Fallback): Require Language: Spanish (dropping the product skill entirely).

2. Injecting the Target Skills in Architect

Before the call hits the queue, Architect must paint the bullseye.

  1. In your Inbound Call Flow, immediately prior to the Transfer to ACD node, add an Update Skills action.
  2. Add all required skills for the absolute perfect match (Ring 1).
    • Add Product_X_Expert (ACD Skill).
    • Add Spanish (Language Skill).
  3. Transfer the call to the IT_Support Queue.

3. Configuring Bullseye Rings in the Queue

Now, configure the queue to degrade those skill requirements over time.

  1. Navigate to Admin > Routing > Queues.
  2. Select the IT_Support queue and navigate to the Routing tab.
  3. Set the Routing Method to Bullseye Routing.
  4. You will now configure the expansion rings.

Configuring Ring 1 (Perfect Match)

  • Time in ring: 120 seconds.
  • Logic: For the first 2 minutes, the system will only alert agents who possess BOTH Spanish and Product_X_Expert.

Configuring Ring 2 (Expansion 1)

  1. Click Add Ring.
  2. Time in ring: 300 seconds (5 minutes).
  3. Skills to remove: Select Product_X_Expert.
  4. Logic: After 2 minutes, the system drops the strict expert requirement. It now alerts any agent who possesses Spanish (and implicitly Product_X_Generalist if they have it, though you don’t explicitly add it here; you simply drop the exclusionary requirement).

Configuring Ring 6 (The Ultimate Fallback)
Genesys Cloud supports up to 6 rings. The final ring has infinite time.

  1. Click Add Ring.
  2. Skills to remove: Select Spanish.
  3. Logic: After 7 total minutes of waiting, the system drops all requirements and alerts any available agent in the IT_Support queue, regardless of language or product knowledge, ensuring the call does not breach maximum SLA.

Validation, Edge Cases & Troubleshooting

Edge Case 1: Proficiency-Based Bullseye

Instead of removing skills (e.g., dropping Product_X_Expert), you can use proficiency ratings (1-5).

  • Setup: In Ring 1, require Product_X_Expert >= 4 (Masters). In Ring 2, lower the requirement to Product_X_Expert >= 2 (Novices).
  • Warning: Do not mix skill removal and proficiency degradation in the same ring blindly. It makes troubleshooting routing logic nearly impossible. Pick one paradigm per queue.

Edge Case 2: Shrinkage and Empty Rings

If Ring 1 has a 5-minute timeout, but all 3 agents with Product_X_Expert called in sick today, callers will sit in Ring 1 for 5 minutes waiting for ghosts, completely ruining SLAs.

  • Troubleshooting: Bullseye Routing is “dumb” regarding staff presence. It will wait the full 5 minutes even if 0 agents are logged in with that skill. To prevent this, use a pre-queue data action in Architect to check GET /api/v2/routing/queues/{id}/estimatedwaittime. If EWT > SLA, strip the expert skill in Architect before transferring to the queue, immediately dropping them into a broader pool.

Validating the Routing Behavior

  1. Open Performance > Workspace > Interactions.
  2. Locate a test interaction that escalated through multiple rings.
  3. Check the Routing Data timeline. You will explicitly see log entries stating: Bullseye Ring 1 Timeout -> Entering Ring 2 -> Removing Skill: Product_X_Expert. If you do not see these logs, your queue is set to Standard Routing, not Bullseye.

Official References