Platform

Queue Management

Skills-based routing over WebSocket — sub-second pickup

Customers wait in a virtual queue, see their position, and get auto-assigned to the longest-idle qualified agent. Configurable overflow, priority, and after-hours rules.

<1s

Median pickup latency

Skills-based

Multi-attribute routing

Per-queue

Independent priority + overflow

Live ETA

Shown to waiting customers

01

Skills-based ACD

Define queues with required skills (language, product, tier, region). Tag agents with the same skills. The router picks the longest-idle agent who matches all required skills for the incoming customer. If no agent matches, the customer overflows to a fallback queue per your policy — or holds with optional callback.

02

What you can configure

Multi-attribute matching, priority customers, time-of-day rules, overflow chains, callback offers, max-wait limits, and abandonment handlers. Every rule is evaluated server-side at pickup time — no race conditions.

  • Multiple queues per tenant, independently configured
  • Skill-based + priority-based matching
  • Overflow chains across queues
  • After-hours fallback — voicemail, callback, IVR-style menus
  • Max-wait and max-queue-depth alarms
  • Per-queue agent allocation caps
  • Live queue depth + abandon rate dashboard
03

What the customer experiences

Position-in-queue, estimated wait, and an optional 'request a callback instead' button — all rendered live in the widget. No third-party callback service to integrate.

04

What the supervisor sees

Live queue depth across all queues, abandon rate trend, longest current wait, agent status board (online / busy / break / offline), and one-click reassignment of waiting customers to a different queue.

SPEC

Technical specifications

Routing
Server-side WebSocket, sub-second match
Storage
Redis Sorted Sets for queue state
Persistence
PostgreSQL for queue config, Redis for runtime
Capacity
Tested to 5,000 concurrent waiters per queue
API
REST (config) + WebSocket (live state)
FAQ

Frequently asked questions

Yes. Tag a customer's session with a priority score (e.g. via a custom widget data field). The router considers priority before idle-time during matching.

Configurable per queue: hold (with live wait estimate), overflow to a fallback queue, offer a callback, send to voicemail, or play a recorded message and end. Most production deployments use a hold + callback offer combination.

Optionally yes — 'agent pull' mode lets an agent claim the next call from a queue manually instead of being auto-assigned. Useful for high-skill or sensitive queues.