Triage every support ticket and draft the first reply automatically
Starts when a help desk or support inbox sends a ticket. Ends with a routed ticket, a deadline and a reply draft, with urgent tickets sent to a person.
Bring the step you are stuck on: 30 minutes, a straight answer and the next step. If you would rather have it built, a reply within 5 working days.
The route
6 steps: 2 rule steps, 1 AI step and 1 person step
Copy, share, discuss
- Copy
Available once this workflow passes its checks.
- Share
todkett.com/automations/customer-support/support-ticket-triage/
- Discuss
Opens when the first workflow passes its checks. No GitHub account? Ask about this by form.
How the library, the tracks and the talk work: the community page.
At a glance
- Starts when
- A help desk, support inbox or chat tool posts tickets to one webhook.
- Input
- Tickets arrive at the webhook, up to 100 per call.
- Output
- A routed ticket, a deadline and a reply draft, with urgent tickets sent to a person.
- AI
- Uses AIClassifying each ticket and drafting the reply.
- Rule checks
- 6 checks around the AI
- Tools
- Webhook from a help desk or support inbox, OpenRouter, Google Sheets
- Last stress test
- No recorded run yet
01Steps
The steps
6 steps: 2 rule steps, 1 AI step and 1 person step. Support is a shared inbox and the urgent ticket is tenth in line. Someone reads every ticket to decide its priority and team, so outages wait behind how-to questions and churn threats get missed.
01Trigger
Ticket arrives
Tickets arrive at the webhook, up to 100 per call.
02Rule step
One item per ticket, capped
One item per ticket; each body capped at 6,000 characters.
03AI step
AI labels and drafts a reply
AI sets category, priority, sentiment and churn mention, and writes a summary, facts to check and a reply draft.
04Rule step
Team and deadline by rule
Fixed rules pick the team and the response deadline, and raise priority for key plans.
05Person step
Urgent tickets to a person
P1 tickets and churn mentions go to a person; every reply draft waits for a person to send it.
06Output
Routed ticket and draft
Routed tickets with their drafts.
Support queue
02Where AI is used
Where AI is used, and what stays with you
AI reads, judges or drafts. Fixed rules check, decide and route. A person approves anything that goes out.
AI step
Step 03Classifying each ticket and drafting the reply.
Tickets are free text. Routing and deadlines stay in code.
What the model reads
- The ticket body, capped at 6,000 characters.
- The written definitions of priorities P1 to P4.
What it never sees or decides
- The team or the deadline: fixed rules set both.
- A key plan's priority: a rule raises it.
- The customer's inbox: every reply draft waits for a person.
Model calls: 1 model call per ticket.
Fixed rules
They check, decide and route. 2 rule steps.
- 02One item per ticket, capped
- 04Team and deadline by rule
A person
- 05P1 tickets and churn mentions go to a person; every reply draft waits for a person to send it.
03Checks
Checks around the AI
What runs before the model, and what checks its answer after, in plain words. Build these into your version too.
- 01Priority definitions P1 to P4 are written into the prompt.
- 02Team and deadline are set by rule.
- 03Priority is raised for key plans by rule.
- 04P1 tickets and churn mentions are escalated to a person by rule.
- 05An unreadable AI answer defaults to P2 with a person's check.
- 06Reply drafts never promise refunds or timelines, and nothing is sent.
04The 8 checks
Stress test
The 8 checks we run on every workflow, and you can run too. A check marked Not needed says why.
- Status
- Not yet0 of 7 that apply have a recorded pass
- Last run
- No recorded run yet
- Model review
- Still open
Checks marked Not yet are still open. 2 of them were seen in the workflow design on 23 Oct 2026; a check shows Passed only once its test run is recorded.
ST7Human sign-off
Status: Not yet
Seen in the workflow design on 23 Oct 2026. The recorded test run is still open.
ST8Rules around the AI
Status: Not yet
Seen in the workflow design on 23 Oct 2026. The recorded test run is still open.
0 passed · 1 not needed · 7 not yet
What we testWe show what we tested, never the test data. The example below is made up.
05Test replay
How to test it
Before you trust a workflow, try to break it. The sample record runs the route step by step. Each break test below gets its own replay once its recorded run exists.
- Record
- TKT-0388
- Data
- Sample run, made-up data
01Ticket arrives
In from a made-up customer of Fernway
Done02One item per ticket, capped
Body within 6,000 characters
Done03AI labels and drafts a reply
Outage, P1, churn mention; summary and reply drafted
Done04Team and deadline by rule
Team and deadline set by rule
Done05Urgent tickets to a person
P1 and a churn mention: held for a person
Waiting for a person06Routed ticket and draft
Routed with its draft; nothing sent
Support queueArrived
TKT-0388 reached the last station: Support queue. Routed with its draft; nothing sent.
Sample run on made-up data. No stress-test run is recorded yet, so this plays the example, not a test. Nothing runs live.
The break tests for this workflow
- ST3Not needed
Same input sent twice
Returns a result only: writes and sends nothing.
- ST2Not yet
Instructions hidden in the input
No recorded run to replay yet.
- ST4Not yet
A connected service fails
No recorded run to replay yet.
- ST8Not yet
The AI returns a broken answer
Seen in the workflow design. No recorded run to replay yet.
06Example
The example, in and out
Made-up data, written to show the steps; no recorded run stands behind it yet. Fernway is a made-up small software company, and its address ends in .example, so it cannot belong to a real business.
InputExample
Example on made-up data: a ticket to Fernway support from a made-up customer: our dashboard has been down since 9 AM and we are thinking about cancelling.
OutputExample
Category outage, priority P1, churn mention yes. Team and deadline set by rule. Escalated to a person because of the P1 and the churn mention. Reply draft saved, not sent.
07Limits
When not to automate this
- It does not send replies, promise refunds or set timelines.
- A small number of tickets a day is faster to sort by hand.
08What you need
What you need to run it
| Tool | Access |
|---|---|
| Your help desk or support inbox | Post each new ticket to one webhook |
| OpenRouter | One API key with a credit limit, for the triage step |
| Google Sheets | Optional: write a test log |
- Where it runs
- In n8n, self-hosted or n8n cloud.
- Model calls
- 1 model call per ticket.
- Running costs
- Yours, paid to each service directly. Scope in writing first. A quote before any build. No prices on the site yet.
How to adapt it
Changes we have made or tested on this workflow.
No tested change is listed yet.
What changed
- ContentFirst version, written from the 23 Oct 2026 catalog.
09FAQ
Questions people ask about it
Does it reply to customers?
No. Reply drafts never leave without a person, and a draft never promises a refund or a timeline.
How are priorities set?
The model labels each ticket using written P1 to P4 definitions. Fixed rules then pick the team and the deadline, raise priority for key plans, and send P1s and churn mentions to a person.
What if the model's answer is unreadable?
The ticket defaults to P2 and a person checks it.
10Related
Where to go next
Use cases
7 steps
6 steps
Can you automate it?
Stuck on your version of this workflow?
Bring the step that does not work. In 30 minutes you get a straight answer and the next step. If you would rather have it built, tell us your tools and the rules you already follow. Scope in writing first. A quote before any build. No prices on the site yet.