Home / Templates / AI Jira Ticket Autopilot

CORTEXFLOW · FREE FOREVER

Automation

AI Jira Ticket Autopilot

October 21, 2026 · Intermediate · Automation

How it works

This template is the idea from the n8n community forum thread “N8n workflow idea: notify your team when an AI agent finishes a Jira ticket” (Oct 21, 2026, 6 replies) — real teams asking for an AI agent that works Jira tickets and tells the team when it’s done. Built end-to-end:

1. STEP 1 — Fetch open Jira tickets — Hourly Trigger fires every hour → CONFIG — Jira + AI holds your host, project key, and thresholds → Build Jira Auth Header base64-encodes your Jira email + API token into a Basic auth header (nothing hardcoded) → Jira — Fetch Open Tickets calls the Jira REST API /rest/api/3/search with JQL project=<key> AND status in ("To Do","Open") ORDER BY created DESC (25 max) → Normalize Tickets flattens each issue to key, summary, description (ADF → plain text), priority, status, reporter, created, and URL.
2. STEP 2 — Dedupe + log NEW — Read Logged Ticket Keys pulls ticket keys already in Google Sheets → Filter New Tickets Only drops anything already seen → Log Tickets — Status NEW appends fresh tickets to the sheet with status NEW.
3. STEP 3 — AI triage with structured output — AI Ticket Triage Agent (model gpt-4o-mini via OpenAI Chat Model — Triage) reads each ticket and returns a strict verdict through Structured Output Parser — Triage: can_resolve (boolean), confidence (0–100), summary, category, resolution_comment, and escalation_reason. The prompt is deliberately conservative — incidents, security issues, billing, customer data, or anything ambiguous gets escalated. → Update Sheet — Status TRIAGED.
4. STEP 4 — Auto-resolve / escalate gate — IF — Auto-Resolve? passes only tickets where can_resolve is true and confidence ≥ your confidence_threshold (75 by default). TRUE branch: Jira — Add Resolution Comment posts the AI’s resolution comment (Jira ADF format) → Jira — Get Transitions fetches the ticket’s available transitions → Pick Done Transition finds the transition named “Done” (and throws a clear error if your Jira workflow names it differently) → Jira — Transition to Done moves the ticket → Jira — Add Resolved Label tags it with your label_resolved label (ai-resolved) → Build Resolved Record rebuilds a clean record → Update Sheet — Status RESOLVED. FALSE branch: Update Sheet — Status ESCALATED records the AI’s escalation reason for a human.
5. STEP 5 — Team digest + Telegram — Merge — Combine Branches collects both branches → Build Team Digest assembles one message: resolved count + escalated count, with per-ticket lines (key, confidence, summary, link) → Telegram — Send Digest pings your team.

Status flow in Sheets: NEW → TRIAGED → RESOLVED / ESCALATED

How this differs from our Shopify AI Support Agent: that template answers incoming customer chats in real time — this one is a back-office autopilot that works your Jira queue on a schedule, resolves what it safely can, and only bothers humans with what it can’t.

Set up steps

You’ll need:

• n8n (self-hosted or cloud)
• Jira Cloud (free) — sign up at atlassian.com → create an API token at Atlassian account → Security → API tokens
• An OpenAI API key — platform.openai.com → API keys
• Google Sheets OAuth2 credential — n8n Credentials → Google Sheets OAuth2 (sheets.google.com)
• A Telegram bot — message @BotFather, send /newbot, copy the token; get your chat ID from @userinfobot

Steps:

1. Import the workflow — in n8n, Workflows → Import from File → select ai-jira-ticket-autopilot.json.
2. CONFIG — Jira + AI — open the Set node and fill in:
   • jira_host — e.g. yourcompany.atlassian.net (no https://)
   • jira_email — the Atlassian account email the API token belongs to
   • jira_api_token — paste your token (replacing PASTE_YOUR_JIRA_API_TOKEN)
   • project_key — e.g. PROJ
   • confidence_threshold — 75 default; raise to 85+ to auto-resolve less, lower to 60 to auto-resolve more
   • spreadsheet_id — your Google Sheet ID (replacing PASTE_YOUR_SPREADSHEET_ID)
   • sheet_name — TicketLog
   • telegram_chat_id — where the team digest lands
   • label_resolved — ai-resolved default; change to match your team’s label convention
3. Google Sheet — create a Google Sheet with a tab named TicketLog and these exact headers: Ticket Key | Summary | Priority | Reporter | Ticket URL | Created | AI Summary | Category | Can Resolve | Confidence | Escalation Reason | Resolution Comment | Status. Put the spreadsheet ID in spreadsheet_id, and select your Google Sheets credential in all five Sheets nodes (Read Logged Ticket Keys, Log Tickets — Status NEW, Update Sheet — Status TRIAGED, Update Sheet — Status ESCALATED, Update Sheet — Status RESOLVED).
4. AI + notifications — select your OpenAI credential in OpenAI Chat Model — Triage (model is gpt-4o-mini by default), select your Telegram credential in Telegram — Send Digest, and activate the workflow.
5. First run — run it once manually: check the sheet flips new tickets NEW → TRIAGED, and that auto-resolved tickets actually show the AI comment + Done status + ai-resolved label in Jira. If Pick Done Transition throws “No Done transition found”, your Jira workflow names the terminal transition differently (e.g. “Closed”) — either rename it in Jira or adjust the name match in that Code node.

Customize it

• Scope the JQL — edit the jql query parameter in Jira — Fetch Open Tickets: add AND priority in (High, Highest) for a VIP queue, or AND labels = bug per team. Keep maxResults under 50 for heavy descriptions.
• Tune the gate — confidence_threshold at 85+ makes the agent very conservative (more escalations); the systemMessage in AI Ticket Triage Agent is where you encode your team’s auto-resolve policy — add your own rules there (e.g. “never auto-resolve tickets tagged production”).
• Faster/slower cadence — change Hourly Trigger to every 15 minutes for a hot queue, or daily for a low-volume project.
• Different transition — if your Done-equivalent is “Closed” or “Resolved”, update the name match in Pick Done Transition.
• Discord instead of Telegram — swap Telegram — Send Digest for a Discord webhook node and reformat the lines in Build Team Digest.
• Per-ticket notifications — move the Telegram node before the merge to ping per resolved ticket instead of one digest.

← All templates