In the quest for streamlining workflows, Jira Automation stands as a powerhouse for operational efficiency. However, even the most refined systems can encounter hiccups. A case in point is the automation rule for reopening tickets upon new comments – a process that, while useful, can backfire when a simple “thank you” from a client triggers an unnecessary cycle of reevaluation.

Here’s a common scenario: A ticket is resolved, the customer leaves a comment, and suddenly, the ticket reopens. This could be an efficient way to ensure no issue gets overlooked, but what if the comment was just a word of appreciation? Suddenly, you’re faced with excess workload, revisiting tickets that didn’t need any further action.

The Use Case

Automating ticket re-opening only when necessary.

Enter the age of AI. By integrating Jira Automation with the prowess of ChatGPT, I have crafted a use case that demonstrates a solution to this quandary. ChatGPT can analyze the context of comments, distinguishing between expressions of gratitude and legitimate requests for further assistance.

Here’s How It Works

  1. A comment on a resolved issue triggers our Jira Automation rule.
  2. Before any re-opening action is taken, ChatGPT evaluates the comment’s content.
  3. If the comment is identified as a simple “thank you,” the ticket remains closed.
  4. However, if the need for additional support is detected, the ticket transitions to “Waiting Triage,” signaling it requires attention.

The Automation Rule

  1. When: Rule is triggered on – The rule is triggered when a comment is made on an issue. This is the event that starts the automation.
  2. Resolution is resolved – This condition checks if the issue’s resolution is set to “Resolved.” If the issue isn’t resolved, the rule will stop here.
  3. Then: Create variable – It creates a smart value variable named {{customerComment}}. This variable will store the content of the comment that triggered the rule.
  4. And: Send web request – The rule sends a POST request to OpenAI. AI evaluates customer comment, if AI identifies that the customer is satisfied with the resolution of their issue, it responds with <THANKS>, indicating that the ticket can remain Closed. Conversely, if the AI determines that the issue remains unresolved or requires additional follow-up, it responds with <REOPEN>.
  5. And: Add value to the audit log – It takes the response from the web request and logs it into the Jira audit log for record-keeping. .
  6. And: Create variable – Another smart value variable is created, this time named {{chatResponse}}. It will store the response ({{webResponse.body.choices.get(0).message.content}}) from the OpenAI that was logged in the previous step.
  7. If: Compare two values – This is a conditional step that checks whether the {{chatResponse}} variable contains the word “REOPEN.”
  8. Then: Transition the issue to – If the condition above is true, the issue is transitioned to the “WAITING TRIAGE” status, which indicates that the issue needs to be reviewed or re-evaluated.

A few screenshots in case you need them:

Especially for the Web Request to OpenAI:

🚩 Don’t forget to hide the value of your Open AI API Key

The (simple) Prompt

{
  "model": "gpt-3.5-turbo",
  "messages": [
    {
      "role": "user",
      "content": "Analyse the provided comment \n {{customerComment}}  \n If the user's comment is expressing gratitude that the ticket was resolved, ChatGPT should respond with '<THANKS>'.  \n\n If the comment indicates that the problem was not solved or more work is needed or that further engagement is required, ChatGPT should respond with '<REOPEN>'."
    }
  ]
}

🦾 Automation and AI in Action 🧠

Step 1: Agent Resolves the ticket.

Step 2: Customer replies saying that it still not working, ChatGPT checks the comment and replies with REOPEN, Automation detects the string and Reopens the ticket.

Step 3: Note that in the comment I made sure to thank as well, just to make it a bit harder for the artificial intelligence. But certainly it wouldn’t be a problem. The ticket remain closed as expected.

This solution exemplifies how AI can optimize your process by reducing unnecessary workload and focusing agents’ efforts where they’re genuinely needed.