Alert your support lead in Slack when Gorgias tickets approach SLA deadline using Claude Cowork

Prerequisites

Claude Desktop with Cowork enabled
Gorgias account email, API key, and domain
Slack Incoming Webhook URL for your SLA warnings channel
Your SLA targets defined: first-response time per channel (e.g., 4 hours for email, 15 minutes for chat)

Overview

Claude Cowork lets you set up a recurring task that polls Gorgias for open tickets, calculates how much SLA time remains on each, and posts a Slack alert for any ticket approaching its deadline. No code to deploy, no webhook infrastructure to maintain — you describe the job in plain language and Claude handles the rest. This is the fastest way to get proactive SLA monitoring running, especially if you don't have an ops engineer on the team.

Step 1: Create a Slack Incoming Webhook

In Slack:

Go to api.slack.com/apps → your app → Incoming Webhooks
Toggle on → Add New Webhook to Workspace
Choose your SLA warnings channel (e.g., #support-sla-warnings)
Copy the webhook URL

Step 2: Document your SLA targets

Before creating the task, write down your SLA targets. Claude needs these to calculate whether a ticket is at risk.

Example:

Channel	First Response Target
Email	4 hours (240 min)
Chat	15 minutes
Contact form	4 hours (240 min)

Also decide your warning threshold — the percentage of SLA time that must elapse before alerting. A good default is 75% (e.g., alert at 3 hours for a 4-hour SLA).

Step 3: Create the Cowork task

Open Claude Desktop → Cowork tab → + New task.

Title: SLA breach warning — Slack alert

Description:

You are a support operations monitor. Check for open Gorgias tickets approaching their SLA first-response deadline and alert the team lead in Slack.
 
Gorgias credentials:
- Domain: your-store
- Auth: HTTP Basic — username: you@company.com, password: YOUR_API_KEY
 
Slack Incoming Webhook: https://hooks.slack.com/services/YOUR/WEBHOOK/URL
 
SLA targets (first response):
- email: 240 minutes
- chat: 15 minutes
- contact-form: 240 minutes
- default: 240 minutes
 
Warning threshold: alert when 75% of SLA time has elapsed.
 
Steps:
 
1. Fetch open tickets:
   GET https://your-store.gorgias.com/api/tickets?status=open&limit=100
 
2. For each ticket, check if it has received an agent reply:
   Look at ticket.messages — if any message has sender.type = "agent" or source.type = "internal", it has been replied to. Skip these tickets.
 
3. For unreplied tickets, calculate SLA status:
   - Get the ticket's created_datetime
   - Determine the SLA target based on ticket.channel (use the targets above)
   - Calculate elapsed minutes = (now - created_datetime) in minutes
   - Calculate remaining minutes = SLA target - elapsed minutes
   - If elapsed >= 75% of the SLA target, this ticket is at risk
 
4. For at-risk tickets, post a Slack message:
   POST [your Slack webhook URL]
   Content-Type: application/json
   Body:
   {
     "text": "SLA Warning — Ticket Approaching Deadline",
     "blocks": [
       {"type": "header", "text": {"type": "plain_text", "text": "SLA Warning — Ticket Approaching Deadline"}},
       {"type": "section", "fields": [
         {"type": "mrkdwn", "text": "*Customer*\n[customer name]"},
         {"type": "mrkdwn", "text": "*Channel*\n[channel]"},
         {"type": "mrkdwn", "text": "*Time Remaining*\n[X min or BREACHED]"},
         {"type": "mrkdwn", "text": "*SLA Target*\n[target] min"}
       ]},
       {"type": "section", "text": {"type": "mrkdwn", "text": "*Subject*\n[ticket subject]"}},
       {"type": "section", "text": {"type": "mrkdwn", "text": "<https://your-store.gorgias.com/app/ticket/[id]|Open in Gorgias>"}}
     ]
   }
   Fill in the bracketed values from the ticket data.
 
5. Do not re-alert on the same ticket ID within the same run or if you've alerted on it in the last hour. Keep a running list of alerted ticket IDs and timestamps in a local JSON file at ~/.sla-alerted.json.
 
6. If multiple tickets are at risk, batch them into a single Slack message with one line per ticket instead of posting separate messages.
 
7. Print a summary: total tickets checked, at-risk tickets found, alerts sent.

Customize the SLA targets in the description

Update the SLA targets in the task description to match your actual policy. If you have per-priority SLAs (e.g., urgent tickets get a 1-hour target), add those as additional rules in step 3.

Step 4: Set the schedule

Frequency: Every 15 minutes during business hours

For teams with 24/7 chat support, run continuously. For email-only teams, business hours is sufficient since email SLAs typically don't count off-hours.

Cowork only runs while Claude Desktop is open

If your machine sleeps or Claude Desktop closes, scheduled tasks are skipped. A ticket approaching its SLA at 2 AM won't trigger an alert until the next run after Claude Desktop opens. If you need round-the-clock SLA monitoring, use the Agent Skill with a cron job on a server.

Step 5: Test it manually

Click Run now and watch the output. A typical run looks like:

Checking open Gorgias tickets for SLA risk...
Fetched 42 open tickets.
 
Ticket #18201 — "Shipping delay on order #4892" (Sarah M.)
  Channel: email | Created: 3h 12m ago | SLA: 240 min | Remaining: 48 min
  -> AT RISK — Slack alert sent
 
Ticket #18195 — "Wrong item received" (James K.)
  Channel: email | Created: 4h 38m ago | SLA: 240 min | Remaining: 0 min
  -> BREACHED — Slack alert sent
 
Ticket #18210 — "Chat about sizing" (Alex P.)
  Channel: chat | Created: 8m ago | SLA: 15 min | Remaining: 7 min
  -> Within SLA (53% elapsed) — skipped
 
Ticket #18215 — "Return label request" (Maria T.)
  Channel: email | Created: 1h 20m ago | SLA: 240 min | Remaining: 160 min
  -> Within SLA (33% elapsed) — skipped
 
Summary: 42 tickets checked, 2 at risk, 2 alerts sent.

Check your Slack channel to verify the message format. If no tickets are at risk, temporarily lower the warning threshold in the description (e.g., change 75% to 10%) to force alerts for testing, then revert.

Step 6: Brief your team lead

Let your support lead know what to expect:

Alerts appear in #support-sla-warnings every 15 minutes (only when tickets are at risk)
Each alert includes customer name, channel, time remaining, and a direct Gorgias link
Multiple at-risk tickets are batched into a single message to reduce noise
The lead should reassign, escalate, or respond to flagged tickets immediately

When to use this approach

You want SLA monitoring running today with no infrastructure
Your team handles fewer than 100 open tickets at a time
You're experimenting with SLA targets and want to adjust thresholds easily by editing the task description

When to switch approaches

You need 24/7 monitoring without depending on your laptop → use the Agent Skill with cron
You need sub-5-minute alerting for chat SLAs → use n8n with a 5-minute schedule trigger
You want to auto-reassign tickets approaching SLA (not just alert) → use n8n, which can chain a Gorgias API call to reassign after alerting

Cost

Claude Desktop with Cowork: included in your Claude subscription
Gorgias API: included in your plan
Slack Incoming Webhooks: free