// Unified Multi-Channel Ticket System

The core problem: support tickets arrive from multiple channels — Intercom live chat, GoHighLevel WhatsApp, email, and SMS. Instead of juggling between platforms, the team operates from a single custom-built interface where every ticket is an opportunity inside GoHighLevel, synced in real time.

// GHL Opportunity Sync

Every ticket exists as a GHL opportunity — stages synced side by side with the app UI

Moving a ticket stage in the app instantly updates the GHL opportunity pipeline

All-in-one storage since GHL has limited cloud storage — the app supplements it for ticket data

// Intercom Chat Embedding

Ticket has Intercom conversation ID — embedded Intercom chat UI loads directly in the app

Agents reply to Intercom tickets without leaving the app — API calls push messages back to Intercom

GHL-origin tickets (WhatsApp/email/SMS) display a different UI — in-progress build to embed GHL conversations tab directly

// Ticket Fields & Filtering

Priority levels: low, medium, high — synced to GHL opportunity custom fields in real time

Categories: onboarding, billing, technical, feature requests — each change writes back to GHL

Assigned owner per ticket — sortable/filterable by status, priority, category, and owner

// AI-Powered Conversation Summaries

Every ticket includes an auto-generated AI summary visible within the Intercom chat UI — designed for leads and team leads who need to review tickets without reading entire conversation threads.

// Summary Generation

AI auto-generates issue description per ticket — what the customer's problem is

Suggested actions included — what the ticket owner should do next

Visible inside the ticket UI for quick lead/manager review

// Token-Efficient Refresh

Summary refreshes every 30 minutes if conversation continues — prevents excessive token usage

If ticket is idle (no new messages), summary stays cached — no unnecessary API calls

// Caching Strategy

AI summary cached in GHL opportunity custom field — not contact-level, opportunity-level

System identifies each ticket's own AI summary via the opportunity field

Custom values used for internal caching — not contact-level, prevents redundant API calls

// Analytics Dashboard

Real-time operational metrics for team leads and management — from individual agent performance to backlog tracking, all powered by the ticket data flowing through the system.

// Ticket Metrics

Total tickets across all channels

Tickets per owner/agent assignment breakdown

Status counts: in-progress, escalated, closed

// Agent Performance

Percentage of closing per agent — close rate relative to total assigned tickets

Average close time per agent — how long from open to resolved

// Daily Operations

Daily metrics: backlog tickets still open and unresolved

Real-time dashboard updates as tickets flow through the system

// AI-Powered Root Cause Analysis

The root cause analysis engine pulls from ALL cached AI summaries across every ticket — transforming individual conversation intelligence into actionable team-level insights. Designed for monthly improvement meetings: "This is where it hurts the most right now."

// At-Risk Client Detection

At-risk clients identified for past 30 days — aggregated from all ticket AI summaries

7-day at-risk view for urgent attention — separate from monthly view

Per-client ticket breakdown: open + closed counts visible in the analysis

// Recurring Issues Breakdown

Product bugs — recurring technical issues across multiple clients

Feature limitations — what users need but the platform doesn't offer

User education + product knowledge gaps — training opportunities surfaced automatically

// Actionable Intelligence

Recommended actions aggregated from all AI summaries — what to prioritize this month

Pie chart: pain point distribution by category — visual breakdown of where support effort concentrates

Cached as custom values — not contact-level, internal custom values to prevent redundant API calls

// Build & Deployment

// Development Stack

Frontend kickstarted with Lovable — rapid UI scaffolding

Vibe coded with Claude Code — AI-assisted development for business logic and API integrations

Private GitHub repository with version control

// Milestones

3 milestones completed and deployed

Used daily by an 8-person live support team

Active development — GHL conversation UI integration in progress

// Architecture Decisions

Custom values for caching — not contact-level, internal custom values to eliminate redundant API calls

Opportunity-level AI summaries — each ticket owns its own cached summary via custom field

30-minute refresh cycle — balances freshness with token cost efficiency

GoHighLevelIntercomLovableClaude CodeGitHubAI SummariesAnalyticsRoot Cause AnalysisCustom SaaSAPI Integration

Hot Prospector