Slack Bot for Customer Support: Ticket Routing Without Zendesk Overhead

Zendesk costs 89 euro per agent per month on the Suite Professional plan. For a 10-person support team that's 10,680 euro per year, before any add-ons. If your support already lives in Slack — most B2B SaaS teams' does — a custom Slack bot can replace 80% of what Zendesk offers for a one-time build cost of about 4,200 euro.

What the bot does

1. Keyword-based channel routing

A customer posts in #support-inbox: "I can't reset my password." The bot pattern-matches "password" or "login" to the #support-account channel and forwards the message with a reference link. The original poster gets a thread reply: "Routed to account team — ETA 2 hours."

The routing config is a flat table in Postgres: regex pattern, target channel, priority, SLA seconds. Editable via a /support-config slash command. No code changes needed when you add a new product area.

2. Auto-assign by expertise

Each agent has a tag list ("billing", "API", "iOS app", "Stripe"). When a ticket lands in a team channel, the bot picks the least-loaded agent whose tags match the ticket topic. Round-robin within the matched subset. Assignment is announced in-thread with a DM to the agent.

Load balancing: the bot counts open tickets per agent in the last 24h. Above 8 open tickets = skipped. Prevents one person from getting buried.

3. SLA timer

Every ticket gets a priority (set by the routing rule or overridden by the agent). The SLA clock starts when the ticket is created and pauses while waiting on the customer. At 75% of the SLA, the bot posts a reminder in-thread. At 100%, it DMs the team lead. At 150%, it escalates to a manager channel.

4. Escalation to Jira

An agent triages a ticket and decides it's a real bug. They click "Escalate to Jira" — a button the bot adds to every ticket thread. A modal asks for project key, component, and severity. On submit, the bot creates the Jira issue via API, links the Slack thread URL in the description, and posts the Jira issue key back into the thread. Engineering takes over from there.

Implementation — 3 weeks

Week 1. Slack app setup, OAuth scopes, message subscription, routing engine. Build the keyword matcher and channel forwarder.

Week 2. Assignment logic, agent tag schema, load balancer, SLA timer with cron job that checks every 5 minutes.

Week 3. Jira integration via API token, Block Kit escalation modal, reporting dashboard (weekly ticket count by team, SLA hit rate, average resolution time).

Cost breakdown

Build: 4,200 euro one-time. Hosting (Render Standard + Postgres): 45 euro per month. Sentry: free tier. Jira API: already paid. Year 1 total: 4,740 euro. Year 2+ ongoing: 540 euro per year.

When NOT to do this

• You have 30+ agents. At that scale Zendesk's reporting, workforce management, and CSAT tooling start to matter. Stick with it.
• You need a customer-facing portal. Customers want a place to log in and see their tickets. The bot only routes internally.
• You're not already on Slack. Don't adopt Slack just to save on Zendesk. The math doesn't work.
• You handle GDPR-sensitive data. PII in Slack messages is awkward to audit. Use a dedicated tool with proper data residency controls.

Comparison

FAQ

Can a Slack bot replace Zendesk?
For teams under 15 support agents handling under 500 tickets per month, yes. The bot covers routing, assignment, SLA tracking, and Jira escalation.

How does keyword routing work?
A regex-pattern config table maps keywords to Slack channels. Editable via a slash command. No code changes when product areas evolve.

What is the SLA timer?
A priority-based timer (P1: 1 hour, P2: 4 hours, P3: 24 hours). The bot posts a 75% reminder, DMs a manager at 100%, escalates at 150%.

What about email tickets?
Use Slack's email-to-channel feature or Front (24 euro per user) to forward emails into a channel. The bot treats them like any other message.