Fast Estimate AI for Jira

User guide

1) Load Jira Issues

Click Load Jira Issues to import issues from your Jira project. You can choose whether issues of hierarchy level 0 (for example Stories, Tasks, etc.) should be imported, or whether issues of hierarchy level 1 (for example Epics) should be included as well. The imported issues then appear as rows in the estimation table.

2) Select the AI model

In the Estimator section, select the AI model you want to use for estimation (for example GPT-4o mini or GPT-5.4 In this standard version of the Fast Estimate AI app, the available AI models are GPT-5.4 and GPT-4o-mini. GPT-5.4 usually delivers good estimates and follow-up questions. GPT-4o-mini can also be useful, but its quality is less consistent. On the other hand, it is about 10 to 50 times cheaper than GPT-5.4 and often costs only a tenth of a cent, or even less, per estimate. It is therefore well suited for testing and for gathering additional information through its questions. These answers can later be reused by GPT-5.4 to generate more precise estimates. More AI models will be available in the Advanced version of Fast Estimate AI.).

3) Estimate the entire project

Before estimating, add a short project description so the AI understands the context. Also decide whether estimates should be returned in points or time. Use the More button to define the developer skills (experience, AI usage, etc.)—these influence how the AI adjusts the estimates. Then click Start AI Estimation to let the AI estimate all issues at once. For each issue, the AI returns min / max / guess values and an Unknown Weight—a measure of how much uncertainty remains.

4) Refine issues with high uncertainty

For issues with a high Unknown Weight, click AI Estimate at the end of the row to open the detail view. The AI has prepared questions for that issue. Each question shows its weight in brackets—indicating how much it contributes to the overall uncertainty.

5) Answer questions and re-estimate

Answer at least some of the questions, then click Send Answers & Re-estimate. The AI re-estimates the issue with the new information, and the Unknown Weight will decrease. Repeat as needed until the uncertainty is acceptable.

All answers are automatically saved and consolidated into a coherent clarifying description under Issue Clarification for each issue—building up a transparent knowledge base that the AI uses in every subsequent re-estimation. This clarification complements and refines the original issue description and helps resolve possible ambiguities. It also serves the developer as a precise specification when implementing the issue.

6) Understand trial usage and own API key usage

• During the trial phase on Jira Cloud, the app includes up to 500 AI-generated estimates for individual Jira issues at no additional charge.
• The quota is calculated per individual Jira issue estimate, not per request, feature invocation, batch, or distinct Jira issue.
• If multiple Jira issues are estimated in one action, each Jira issue estimated counts separately. Re-estimating the same Jira issue counts again.
• After the free trial quota is exhausted, you must provide your own OpenAI API key for continued AI estimation usage.

7) Enter your own estimates

Users can also enter their own estimates in two different ways.

First way: In the main view, select Me in the Estimator. You can then directly enter your own estimate in the upper table for each Jira issue or work package: the optimistic case (Min), the pessimistic case (Max), and your own best estimate based on the currently available information (Guess). In addition, you can enter an Unknown % value that expresses how much additional effort may be hidden behind this issue although it is not yet clear from the current description. This percentage is measured relative to the effort that is already apparent from the description. Therefore, the value can be well above 100%, sometimes even 300% or more.

Second way: Use the AI Estimate button in an individual row of the table. When you click it, a dialog opens in which you can also enter Min, Max, Best Guess, and Unknown %. In the same dialog, you can see the Issue Description from Jira and the Issue Clarification that the AIs automatically create based on the user’s answers.

Whenever an AI estimate is made and important questions are still open, the AI asks follow-up questions. The user can answer them directly in the input field at the bottom of the dialog. In this dialog, the user may enter their own estimate before reading the AI results in order to compare their own judgment with the AI estimates. Or they can estimate the issue themselves afterwards, once they have reviewed the AI questions, the answers from other contributors if available, and the consolidated clarification. This clarification is also very useful for developers who later need to implement the issue.

Installation guide

1. In Jira Cloud, open Apps → Find new apps (or open Marketplace in your browser).
2. Search for Fast Estimate AI and click Get it now.
3. Approve requested permissions.
4. Open a project and navigate to Fast Estimate AI in the project menu.

FAQ

Support