Customer service / Beginner tutorial

New to customer service? Start here

Complete eight guided steps in about 15 minutes to configure, connect, serve, and review. Every step shows where to go, what to do, and how to verify success.

Technical integration docs
Admin
Create domains and sites, add people, set rules
Open platform
Get AppKey, embed code, debug entry
App
Receive, reply, transfer, and close sessions

Get running in 15 minutes

Learn the three surfaces, then complete eight steps

Customer service is not a single page: Admin owns rules, Open Platform owns web integration, and the App is where agents serve visitors. Beginners often get stuck by looking in the wrong surface.

01

Configure

Admin

Create domains and manage sites, members, and service policies.Admin → Customer service
02

Integrate

Open Platform

Get the public AppKey and connect a widget or standalone page.Open Platform → Web CS
03

Serve

App

Go online, claim requests, reply, transfer, and close conversations.App → Customer service desk
Before you start
  • An administrator account for Admin
  • At least one existing App user for the agent
  • The HTTPS origin of the target website
  • A second device or private window for the visitor

Step 1 / 8

Create the boundary for sites, people, and rules

A customer-service domain is an independent team boundary. Start with one domain per brand or business line.
Who does this
Super admin / domain admin
Where to go
Admin → Customer service → Domains
Estimated time
2 minutes

Follow these actions

  1. 1

    Click New domain.

  2. 2

    Enter a recognizable name, welcome message, and assignment strategy.

  3. 3

    Save and open the domain details.

Verify completion

  • A stable domain code is generated.
  • The domain status is Enabled.

Common mistake: A service domain is not a DNS domain. It groups service sites, people, and permissions.

How one request moves

Understand the service loop first

Visitor entry, agent handling, and administration are one connected workflow. An incomplete stage affects the final service result.

  1. 01Connect a site

    Embed the widget or publish a standalone service link.

  2. 02Identify the visitor

    Keep identity continuity with externalUserId, fingerprint, or local visitor ID.

  3. 03Queue and assign

    Route requests through manual pickup or round-robin assignment.

  4. 04Handle the request

    Claim the conversation, review context, and reply efficiently.

  5. 05Collaborate

    Transfer, escalate, or create a ticket for longer-running work.

  6. 06Close and rate

    Record the outcome and invite the visitor to rate the service.

  7. 07Review and improve

    Analyze volume, first response, transfers, satisfaction, and SLA.

Role-based tasks

Choose a role and focus on the work that matters

Each role uses a different surface. Every step shows where to act and how to verify completion.

Administrator

Configure domains, people, and service policies first

Administrators decide where visitors enter, who can receive them, when service is available, and what happens after a timeout.

Before you start
Super administrator or customer-service domain administrator access.
Primary entry
Admin → Customer service
  1. 1

    Create the first access domain

    Create a domain with a name, welcome message, and default assignment strategy.

    Where
    Customer service → Domains
    Completion check
    The domain is active and has a stable domain code.
  2. 2

    Create access sites

    Create separate sites for the main site, store, or help center and enter every allowed origin.

    Where
    Domain details → Access sites
    Completion check
    A public AppKey exists and the target origin is allowlisted.
  3. 3

    Add members and roles

    Add agents, viewers, or domain admins and set concurrency limits.

    Where
    Domain details → Members
    Completion check
    Agents can open the App desk and switch to receiving.
  4. 4

    Configure hours and timeouts

    Set working hours, offline messages, unattended replies, reassignment, auto-close, and SLA thresholds.

    Where
    Access site → Policy settings
    Completion check
    Business-hour messages, timeout actions, and SLA tracking behave as expected.
  5. 5

    Prepare service tools

    Maintain quick replies and verify permissions for tickets, escalations, and reports.

    Where
    Customer service → Quick replies / Tickets / Escalations
    Completion check
    Agents can search and use active quick replies in the desk.

Capability map

Six capability areas show what the system covers

This table explains business outcomes. Use the technical docs for parameters, code, and event payloads.

AreaIncluded functionsOutcome
Visitor entry
  • CDN / npm embed
  • Standalone links
  • Origin allowlist
  • Visitor identity
  • Theme and locale

Unify traffic from websites, H5, SMS, email, and QR codes into one domain.

Routing
  • Waiting queue
  • Manual / round robin
  • Receiving status
  • Concurrency
  • Timeout reassignment

Control who receives work, how much they receive, and what happens when no one responds.

Conversation handling
  • Quick replies
  • Media and files
  • Visitor profile
  • Tags
  • Internal notes
  • Risk blocking

Help agents understand context quickly and preserve it for the next collaborator.

Team collaboration
  • Transfer
  • Escalation
  • Ticket creation
  • Ticket reassignment
  • Public / internal comments

Move realtime work fast and track complex issues across time and people.

Service governance
  • Member roles
  • Working hours
  • Offline messages
  • Unattended replies
  • Outcomes and close reasons

Turn service policy into system behavior instead of relying on memory.

Analytics
  • Volume and first response
  • Sessions and transfers
  • Agent KPIs
  • Satisfaction
  • SLA
  • CSV export

Drill down by agent, site, and time range to find concrete improvements.

Status reference

A status tells you who should act next

Conversations and tickets use separate lifecycles: conversations handle realtime service, while tickets track longer-running work.

Conversation states

The lifecycle of one realtime visitor request.

  1. Waiting

    The visitor is queued for claim or assignment.

  2. Active

    A current owner can communicate with the visitor.

  3. Transferred

    Ownership changed and the new agent continues with existing context.

  4. Reopened

    A visitor returned after closure and started a new service stage.

  5. Closed

    Realtime service ended and results feed reporting and ratings.

Ticket states

The lifecycle of a problem that needs continued work.

  1. Open

    The ticket exists but has not been claimed.

  2. In progress

    An assignee is actively working on it.

  3. Pending customer

    More information or confirmation is needed from the customer.

  4. Resolved

    A solution is available and awaits final closure.

  5. Closed / Reopened

    The issue is archived or reopened when more work is required.

Next step

Complete role setup before code integration

Create the domain and site first, confirm agents can receive, and only then integrate the webpage. That sequence makes failures easy to locate.