✓ Median first-response time down 60%
✓ 92% of incoming messages auto-routed and tagged
✓ Zero missed tickets after going live
Before → after
| Metric | Before | After |
|---|---|---|
| Median first-response time | 3h 20m | 1h 18m |
| Messages auto-routed & tagged | 0% | 92% |
| Tickets missed per week | ~40 | 0 |
A fast-growing creator platform ran its community on Discord and its support queue in a separate help desk. Moderators switched between tools, threads got orphaned, and the same question was answered twice in two places.
The challenge
- Community questions and support tickets lived in disconnected systems.
- No way to see a user’s full history across Discord and support.
- Manual triage meant slow first responses and missed messages.
- Moderators wanted a single inbox without leaving Discord.
What we built
- Discord bot listens to forums, threads, and DMs, ingesting every message in real time.
- Unified CRM stores members, threads, and tickets in Postgres with a shared timeline per user.
- Auto-triage classifies and tags each message (bug, billing, feature, general) and routes it to the right queue.
- Slack bridge pings the support team with a digest of anything unresolved for over an hour.
- MCP server lets the team query the CRM from an AI assistant (“show me every open billing ticket from this week”).
The results
First-response time dropped 60%, 92% of messages are now auto-routed with correct tags, and the team hasn’t missed a ticket since launch. The same pattern, bot plus CRM plus MCP, is now being extended to their Telegram community.
The takeaway
Community and support are the same problem viewed from two angles. One system with a shared timeline beats two disconnected tools every time.
Stack Discord APIPostgresCloud RunSlackMCP
“
We finally see every community thread and support ticket in one place, with context.