All Departments

Customer Support

OperationsPhase 2

Lead: CS Orchestrator Agent

10 agents10 planned

Architecture Pattern

View full details →

Customer queries are classified and routed to specialized handlers (account, payment, game, general). Each handler follows a resolution chain. Proactive Support and Knowledge Base agents work in parallel to deflect tickets before they arrive.

Evaluation
6 metrics defined
Error Handling
4 scenarios covered
Self-Improvement
3 human checkpoints
Phase 2: Core Operations

Core Operations. Automated CS via Intercom handles player queries and reduces manual workload. Depends on Phase 1 player data and risk flags.

Tools

IntercomBack Office APILinearSlack API

Goals

Automate support operationsResolve tickets fastReduce human escalationSelf-service knowledge baseProactive issue detectionIn-app notifications & broadcasts

Agent Sizing Rationale

8 agents: 1 orchestrator + 3 support specialists (odd for escalation voting) + 1 quality + 1 knowledge base + 1 proactive support + 1 station message. 3-agent panel votes on whether to auto-resolve or escalate complex tickets.

Support (3-agent resolution panel)

AgentDescriptionComplexityRolesStatusActions
Live Chat AgentHandles real-time player queries via Intercom. Resolves common issues (balance, bonus, verification). Uses conversation history and player context for personalized responses.medium
executor
Ticket Resolution AgentProcesses async support tickets, categorizes issues, and applies resolution playbooks. Handles email and form submissions.medium
executorcommunicator
Escalation & Complaints AgentHandles complex complaints, VIP issues, and regulatory-sensitive queries. Manages formal complaint procedures and tracks resolution.medium
monitor

Quality, Knowledge & Proactive Support

AgentDescriptionComplexityRolesStatusActions
CS Quality & CSAT AgentMonitors response quality, CSAT scores, resolution times, and identifies training gaps. Generates weekly quality reports and flags underperforming resolution patterns.medium
monitoranalystgenerator
Knowledge Base & FAQ AgentMaintains the self-service help center — creates and updates FAQ articles, how-to guides, and troubleshooting flows. Analyzes top ticket categories to proactively create content that deflects future tickets.medium
analystgeneratorcommunicator
Proactive Support AgentDetects player issues before they report them — failed deposits, stuck withdrawals, KYC delays, game errors. Sends proactive messages via Intercom or Station Message to affected players with resolution steps.medium
monitorexecutoranalyst
Station Message & Notifications AgentManages in-app station messages and broadcast notifications. Sends targeted announcements (maintenance, promotions, system updates) to specific user segments or all users. Handles message scheduling, read tracking, and priority levels.simple
monitorexecutorcommunicator
CS Context Enrichment AgentInjects real-time player data into the live chat sidebar when a conversation starts — VIP level, GGR, total bonuses, KYC status, CCF, wallet connection method, active tags, and recent transaction history. Eliminates the need for CS agents to manually look up player profiles.medium
monitoranalyst
Knowledge Base Curator AgentMonitors AI chatbot responses for accuracy and language matching. Flags incorrect answers, auto-updates the knowledge base from Notion documentation, ensures bot responds in the user's detected language, and generates weekly quality reports posted to #cs-improvements.medium
monitorgenerator
CS Orchestrator AgentRoutes tickets to the right agent, manages workload, maintains knowledge base, and reports CS KPIs (CSAT, FCR, response time, escalation rate).complex
orchestratorexecutoranalyst