AI Generator for AI-powered release notes
For ARNR Server-DC app
If you are using ARNR on Jira Server/DC, you will have to connect your Open AI account from the App Settings page to start using this feature. For Jira cloud users, no integration is required.
Once connected, select the Open AI connection in the dropdown and follow the below steps for AI-generated release notes.
Want to connect your own Open AI account for Cloud ARNR app?
Raise a request on out Service desk for early access to the advanced edition of the app. Read ARNR Advanced editions and Free vs Paid apps for more details.
Using the new AI Generator feature, you can easily generate release notes through AI by simply selecting an input Jira field. The system will read the contents of the selected field, summarize or rephrase them, and provide refined and meaningful release notes as an output. You can also choose the tone of your release notes.
Enable the ‘AI Generator’ feature
Your Jira admin shall have to enable this app setting -Allow release notes generation through AI. Admin can enable it from the ARNR configurations screen that is available within the ‘Manage apps’ menu. Know more.
This document is in progress to be updated with the latest UI changes!
Navigation
Navigate to the ARNR from the relevant Project’s in-project menu
Click on AI Generator tab
Using prompts
Deafult prompt
When using deafult prompt, you need to provide details in the given fields to execute AI-generated content in the given issues. Below are the fields and their details.
Currently ARNR supports multiline text and single-line text Jira issue/custom fields for AI-generated release notes.
Custom prompt
Custom prompt allows you to provide a detailed prompt to the AI Generator so it can provide you results based on the given prompt. In the custom prompt, all the fields are applicable except Action, Tone, Maximum words, Output language as these details can be added to the custom prompt itself.
Fields in the AI Generator
Field | Description | Applicable for |
|---|---|---|
Action | Define whether you wish to summarise or rephrase the content | Default prompt |
Tone | Define the tone from the given options like Professional, Casual, Humorous | Default prompt |
Maximum words | Specify the maximum number of words in the output | Default prompt |
Output language | Specify output language. Current options are - English (UK and US), French, Dutch, Polish, Spanish, and German. | Default prompt |
Where should the AI fetch data from?
| Input Jira issue field - Jira issue field that will provide the input to the AI Generator | Default and Custom prompt |
Where to store release notes? | With this field, there are two options -
| Default and Custom prompt |
JQL to fetch issues for which AI-assisted release notes are needed (Mandatory) | JQL to fetch issues that you want to update with the output suggested by the AI Generator | Default and Custom prompt |
Hide issues which have "Input Jira issue field" empty | Check this box if you wish to hide issues with no input data in the selected Jira issues field | Default and Custom prompt |
Get release notes suggestions | Once all the above information is updated. Click on this button to generate the suggestions.
| Default and Custom prompt |
Step-by-Step Workflow
1. Configure
Select your prompt type (Default or Custom). If you select the default prompt, select relevant options from defaul dropdowns. If you select a custom prompt, add your prompt for the AI generator with detailed instructions.
Choose your Input Field (e.g., Description), and define your Output Field to store AI-generated conten. Click Continue.
2. Fetch and Generate
Fetch Issues: Enter your JQL (e.g.,
project = RR AND fixVersion = "v1.1") and click Fetch Issues.Use Hide issues which have all "<input field>" empty checkboxes to avoid processing issues with the empty input field you have selected.
Generate Results: Select the desired issues from the list and click Generate Results. The AI will process the descriptions based on your prompt settings.
Review & Edit: The generated snippets appear in the AI-generated content column. You can manually refine the text by clicking on the pencil icon in the Actions column. If needed, use the Restore option to revert to the original AI suggestion.
3. Review and Save
Summary: Review the count of updated, modified, and skipped issues.
Notifications: Enable the Update notification option to send a Jira email to Watchers, Assignees, etc. once the field is updated.
Save to Jira: Click Save to Jira to commit the changes. A success message will confirm the data is stored.
If needed, you can use Back button to make any changes or re-generate content.
Use Case
A Product Manager wants to convert internal technical descriptions into public-facing notes. They select the internal Description as the input and a custom field, Public Release Notes, as the output. By using a Professional tone and summarizing actions, the AI generates concise snippets for all selected issues, which are then automatically saved to the public field for use in ARNR templates.
Troubleshoot errors
Update issue action can fail in the following cases
When the user lacks permission to update the output field
When the user lacks permission to disable email update notifications
Click on check box Send mail for this update
Only the top 100 issues are displayed as maximum 100 issues are supported at a time.
Change the JQL so it fetches issues in batches of fewer than 100 issues at a time.
Read Actions to create AI-generated release notes using rule action.