Many AI tools promise that you can simply describe what you want and get a working result. And in many cases, that’s true. For example, I take the tool I'm working on, AI Apps Builder for Jira. Using this app, you can explain the Jira app you need in plain language, and the AI generates a Forge app for you.

However, one thing becomes clear very quickly: the more specific your description is, the better the result will be.

While experimenting with AI Apps Builder for Jira, certain patterns started to emerge. Some prompts consistently produced clear, usable apps, while others led to results that needed refinement. The difference wasn’t the AI itself — it was how the request was written.

In this article, we’ll look at those patterns and explore practical ways to write better prompts when building custom Jira apps, dashboards, and reports.

Best Practices for Writing Effective AI Prompts

1. Start With a Clear Goal

Before writing a prompt, it helps to answer a simple question: What are you trying to achieve?

For AI Apps Builder, that usually means defining:

• What do you want to build?

• Where it should appear in Jira

• What outcome should the app produce?

Weak prompt

Create a dashboard gadget that tracks sprint health.

Improved prompt

Create a dashboard gadget that tracks sprint health for a selected project and sprint. Include a burndown chart and a summary of completion status in Jira style.

Starting with a clear goal helps the AI generate the correct structure from the beginning.

2. Define the App Type First

Jira apps can appear in different locations:

• Dashboard gadgets

• Issue panels

• Global pages

Each of these modules behaves differently. If the prompt specifies the module type at the beginning, it reduces ambiguity.

Weak prompt

Show sprint progress.

Better prompt

Create an Issue Panel for Jira that shows an overview of the current sprint.

Naming the module helps the system select the right Forge architecture immediately.

3. Be Explicit About Data Scope

AI systems can’t infer details that aren’t provided.

Whenever possible, define:

• time range

• project scope

• statuses

• issue types

Weak prompt

Show sprint velocity.

Better prompt

Show velocity for the last 3 sprints, plotting committed vs completed story points and calculating average velocity.

A few extra details can turn a vague request into a clear, data-driven custom report.

4. Define Business Logic Clearly

Sometimes prompts include concepts that seem obvious to humans but are unclear to machines.

For example: Show high-risk issues. What qualifies as “high risk”? It helps to define the logic explicitly.

Example

Identify issues as high risk if they are flagged, overdue, or due within the next 7 days.

This allows the AI to understand exactly how Jira data should be interpreted.

5. Describe Layout and UI Expectations

AI Apps Builder generates both logic and interface, so it helps to describe the UI.

Useful elements to include:

• cards

• charts

• tables

• color rules

• layout structure

Example

Display workload as horizontal bars, highlight overdue issues in red, and use badges for status indicators.

Small design hints can dramatically improve usability.

6. Add Interactivity

Interactive elements make dashboards more useful. Consider adding instructions such as:

• dropdown filters

• multi-select filters

• click actions

• expandable sections

Example

Show workload per user and add multi-select filters for project and assignee.

This allows users to explore the data rather than view a static report.

7. Use a Structured Prompt Template

A simple structure works well with AI Apps Builder.

Create a [dashboard gadget / issue panel / widget] that [main purpose].

Data Scope:
[projects]
[time range]
[statuses]
[issue types]

Display:
[charts / summary metrics / tables]

Filters:
[dropdowns / date range / multi-select]

Logic:
[definitions of conditions]

UI:
[Jira style / layout preference]

Example Prompt

Create a dashboard gadget that analyzes sprint performance.

Data Scope:
Selected project
Last 3 completed sprints

Display:
Velocity trend chart (committed vs completed story points)
Average velocity
Percentage change compared to the previous sprint

Filters:
Project dropdown

Visual Indicators:
Upward trend in green
Downward trend in red

UI:
Use Jira style for UI.

Using a structure like this helps the AI generate more consistent results.

8. Do a Quick Prompt Check Before Generating

Before clicking Generate, a quick checklist can help:

• Did I define what I want to build?

• Did I specify the Jira module type?

• Did I define time range and data scope?

• Did I clarify the logic?

• Did I describe the UI expectations?

If most answers are yes, the prompt is likely ready.

Writing a good AI prompt is surprisingly similar to writing a clear product requirement. The more specific you are about the goal, data scope, and structure, the better the generated solution will be. Tools like AI Apps Builder for Jira make it possible to experiment with custom Jira apps without writing code. But the real skill becomes learning how to describe what you want clearly.

For teams exploring AI-driven development, no-code tools, or Forge apps, prompt writing is quickly becoming an important skill. And like any skill, it improves with practice.

Try AI Apps Builder and Support the Launch

If this idea resonates with you, we’d love to hear your thoughts. We’ve launched AI Apps Builder for Jira on Product Hunt, and the goal is simple: help Jira teams build custom apps faster using natural language.

You can check it out here: https://www.producthunt.com/products/ai-apps-builder-for-jira/

And if you’ve ever thought,

“I wish Jira had a report for this…”

This might be a good moment to try building it.