Knowledge Base / CALL_CENTER

Genesys Cloud

Vendor: Genesys · All

Genesys Cloud CX is Genesys's cloud-native (SaaS) contact center and experience orchestration platform. It handles inbound and outbound voice plus digital channels, with Architect IVR / bot flows for self-service, ACD routing to agents, and built-in workforce engagement. All interaction activity is exposed through the Analytics API as conversation, session and segment records.

Call Handling (CALL_HANDLING)

Inbound voice contact handling, from call arrival through IVR self-service, ACD queueing, agent handling and after-call wrap-up. The case is the individual call / contact.

Activities

• Call Received - Call arrives; Conversation.conversationStart and first Session open.
• IVR Menu Presented - segmentType=ivr; the Architect flow presents the menu (Flow.entryReason).
• Menu Selection Captured - Caller DTMF/speech selection captured within the ivr segment (Flow).
• Self-Service Attempted - The flow executes the caller's request (data dip / transaction).
• Self-Service Resolved - Flow.exitReason indicates the call was contained without an agent.
• Queued to ACD - Segment gains queueId; the call is offered to an ACD queue (purpose=acd).
• Wait in Queue - interact-in-queue segment duration before an agent is alerted.
• Caller Abandoned - Caller disconnects while in queue; disconnectType=peer before Agent Answered.
• Agent Alerted - segmentType=alert; a routed agent is being offered the call.
• Agent Answered - segmentType=interact with selectedAgentId; the agent is talking to the caller.
• Call Held - segmentType=hold; the agent places the caller on hold.
• Call Transferred - segmentType=transfer; the call is handed to another queue/agent.
• After-Call Work / Wrap-Up - segmentType=wrapup; the agent sets Segment.wrapUpCode.
• Call Ended - Final disconnectType set and Conversation.conversationEnd stamped.
• Voicemail Left - segmentType=voicemail; caller leaves a message instead of reaching an agent.
• Callback Scheduled - segmentType=scheduled; a callback is booked to be dialed later.

Case attributes

• Media Type (string)
• Originating Direction (string)
• Disconnect Type (string)
• ANI (string)
• DNIS (string)
• Wrap-Up Code (string)
• Handling Agent (string)
• Conversation ID (string)
• Division (string)
• Queue (string)
• Flow Exit Reason (string)
• Flow Name (string)

Process-mining benefits (historical)

• See your true IVR containment rate [Service]: Measure how many calls the IVR resolves on its own versus how many fall through to an agent - the honest containment number, by flow.
• Find the flow steps where callers opt out to an agent [Efficiency]: Pinpoint the exact IVR menu options where callers give up on self-service and ask for a person, so the flow can be fixed where it leaks.
• Quantify the wait between IVR and a live agent [Efficiency]: Measure how long callers sit between leaving the IVR and an agent answering, and which queues and times drive the wait.
• Understand where and when callers abandon [Service]: Rank queues and time windows by how often callers hang up before an agent answers, so service gaps are obvious.
• Explain handle-time variation [Cost]: Compare talk, hold, and after-call-work time across agents, queues, and wrap-up codes to see what really drives long calls - and where coaching pays back.
• Expose transfer and re-queue loops [Efficiency]: Surface calls bounced between queues and agents before they were resolved - the classic first-contact-resolution failure.

Process-mining benefits (action)

• Alert when containment drops on a flow [Service]: Watch self-service containment per flow and raise an alert when it slips, so a broken menu or data dip is caught the day it starts leaking to agents.
• Route known repeat callers to priority handling [Service]: Recognise a caller who has phoned recently and route them straight to priority handling instead of the back of the queue.
• Trigger overflow when handoff delay climbs [Efficiency]: Watch the live IVR-to-agent wait and trigger overflow routing or extra staff the moment a queue's handoff delay climbs past target.
• Offer a callback to calls at risk of abandoning [Service]: Spot calls waiting past the point where abandonment spikes and offer a callback before the caller hangs up.
• Catch stuck or excessive wrap-up in real time [Cost]: Alert when after-call work runs far beyond the norm, keeping agents available and wrap-up honest.
• Work a transfer-loop worklist [Efficiency]: Give supervisors a daily list of calls transferred more than once so the routing or skill gap behind them gets fixed at the source.

Data tables

• Conversation - Analytics API conversation record: one per interaction, from conversationStart to conversationEnd, with originatingDirection and the list of participants. The case is the conversationId.
• Segment - A state within a session with a start and an end (segmentType: ivr, interact, alert, delay, hold, wrapup). The event grain used to reconstruct the call's path.
• Session - A participant's activity on one media type (voice). Carries ani, dnis, direction and mediaType, and may carry Architect flow data; parent of segments.
• Flow - The Architect flow (IVR / bot) that processed the interaction, captured in the session's flow data: flowId, flowName, entry and exit reason. The self-service navigation record.
• Participant - A party in the conversation, tagged with a purpose (customer, ivr, acd, agent). Each participant owns one or more sessions.
• Queue - Queue
• User (Agent) - User (Agent)
• Wrap-Up Code - Wrap-Up Code

Service/Case Management (SERVICE)

Support tickets, SLAs, resolution

Activities

• Receive Contact - Service contact opens; Conversation.conversationStart (or a routed Workitem).
• Identify Customer - External Contact resolved and linked to the interaction (participant.externalContactId).
• Categorize Request - Routing skills/queue classify the request; Segment.requestedRoutingSkillIds / queueId.
• Assign to Agent - Interaction/workitem routed and connected to an agent; Segment.selectedAgentId / Workitem.assigneeId.
• Handle Interaction - Agent works the request; segmentType=interact talk time.
• Consult / Escalate - Hand to a specialist/supervisor; segmentType=transfer.
• Create Workitem - Open a Task Management workitem for follow-up work (requires Task Management); Workitem.dateCreated.
• Schedule Callback - Book a follow-up callback; Callback.callbackScheduledTime / segmentType=scheduled.
• Apply Disposition (Wrap-Up) - Agent records the outcome; Segment.wrapUpCode during segmentType=wrapup.
• Resolve Request - Request resolved on the interaction; wrapUpCode/disconnectType indicating resolution.
• Quality Evaluation - QM evaluation scored for the interaction (requires Quality Management); Evaluation.totalScore.
• Close Workitem - Follow-up workitem reaches a terminal status; Workitem.statusId / dateClosed.
• Repeat Contact - Same External Contact returns within a window - a reopen/unresolved signal.

Case attributes

• QM Evaluation Score (number)
• Workitem Type (string)
• Media Type (string)
• External Contact (string)
• Resolution Outcome (string)
• Wrap-Up Code (string)
• Callback Scheduled (boolean)
• Escalated (boolean)
• Handling Agent (string)
• Conversation ID (string)
• Queue (string)
• Skills Requested (string)

Process-mining benefits (historical)

• Measure real first-contact resolution [Service]: See how often a request is genuinely resolved on the first contact versus how often it spawns a follow-up workitem or a repeat contact - the honest FCR number, not the wrap-up code's optimistic one.
• Trace repeat contacts to their root [Service]: Follow the same customer coming back for the same issue and group the request types that drive avoidable repeat contacts.
• Expose where and why work escalates [Efficiency]: Rank the request categories and queues that escalate to specialists most often, so the routing or knowledge gap behind them can be closed.
• Find aging and stuck follow-up work [Efficiency]: Measure how long follow-up workitems sit between creation and closure, and which types and queues let them age, so backlog stops hiding.
• Audit disposition accuracy [Compliance]: Compare the wrap-up code agents apply against what actually happened (reopen, escalation, callback), exposing mis-disposition that corrupts reporting.
• Link quality scores to outcomes [Cost]: Test whether low quality-evaluation scores line up with escalations and reopens, so coaching targets the handling that actually costs rework.

Process-mining benefits (action)

• Work an aging-workitem list daily [Efficiency]: Give each queue and agent a live list of open follow-up workitems past their target age, so back-office work is cleared before it breaches.
• Alert on a returning customer [Service]: Flag when a known customer contacts again within a short window so the agent handles it as a continuation - not a cold new ticket - and the repeat is counted.
• Route escalation-prone requests to a fix [Efficiency]: Feed the request types with the highest escalation rate into training and process changes, cutting the transfers before they happen.
• Trigger on missing or invalid disposition [Compliance]: Flag contacts closed with no wrap-up code or an outcome that contradicts a later reopen, keeping disposition data trustworthy.
• Chase scheduled callbacks to completion [Service]: Keep a live list of scheduled callbacks so every promised follow-up is actually made and closed.
• Target quality reviews where they matter [Cost]: Auto-select interactions that escalated or were reopened for quality evaluation, focusing scarce QM effort on the handling most likely to be at fault.

Data tables

• Conversation - Analytics API conversation record: one per interaction, from conversationStart to conversationEnd, with originatingDirection and the list of participants. The case is the conversationId.
• External Contact - The customer/case subject. Ties a service request to a person and to prior contacts (repeat-contact detection).
• Segment - A state within a session with a start and an end (segmentType: ivr, interact, alert, delay, hold, wrapup). The event grain used to reconstruct the call's path.
• Workitem - Task Management work item - a native case/task the agent opens or is routed. Carries status, type, queue, assignee, and the external contact.
• Callback - A scheduled callback follow-up interaction. Requires callback to be configured.
• Evaluation - Quality Management evaluation of a handled interaction. Requires Quality Management; scores/coaching are extractable only when evaluations are done.
• Queue - Queue
• User (Agent) - User (Agent)
• Wrap-Up Code - Wrap-Up Code

Explore this interactively in the mindzie Knowledge Base.