AI-powered relative estimation and forecasting for Jira Cloud
Relativity is an AI-powered relative estimation and forecasting app for Jira Cloud. It analyzes your completed epics to discover natural effort bands, forecast delivery timelines, and help your team make data-driven planning decisions.
Relativity uses AI-powered pattern recognition to detect natural effort bands — groupings of similar-sized epics that emerge from your actual delivery history.
How to use:
issuetype = Epic AND status = Done AND resolved >= -6mControls:
Click “Compare” on any epic card to add it to the comparison panel. Compare epics across bands with aggregate stats:
Plan future work by exploring how different story point estimates would track against your historical data.
How to use:
Standard tier adds: 5 most similar past epics with actual cycle time, velocity, and complexity data.
A bird’s-eye view of your delivery intelligence:
Get a live AI-powered delivery prediction for any epic — in-progress or completed.
How to use:
Save analysis results to revisit later. Load any saved analysis to see the full results without re-running the query.
Track how effort bands shift over time across multiple saved analyses. Stable lines suggest consistent sizing patterns. Converging lines indicate improving estimation consistency.
Set up automated weekly or monthly analysis runs on saved JQL queries. Results are saved automatically for trend tracking.
Natural groupings of epic sizes identified by AI-powered analysis. Each band represents a cluster of epics with similar story point totals. Bands are not manually defined — they emerge from your actual data.
The number of days from when the first child issue starts work to when the last child issue completes. Calculated from Jira status transitions (when issues move to In Progress and Done).
A 0–100 score computed from four signals: number of child issues, comments, links, and commits. Weights are adaptive — signals with more variance in your dataset get higher weight because they differentiate complexity more effectively.
Story points delivered per day. Calculated as total story points divided by Epic Cycle Time. Higher velocity means faster delivery throughput.
How consistent both story point sizing and delivery time are within a band. Uses adaptive AI weights — the signal (sizing spread vs cycle time spread) that varies more gets more weight. High predictability means you can reliably predict both size and delivery time for epics in that band.
Completed epics that took longer AND were more complex than typical for their band. These are worth investigating to understand what caused the deviation — scope creep, blockers, dependencies, or other factors.
An AI-weighted composite score (0–100) measuring overall delivery health across four dimensions: predictability, risk rate, velocity consistency, and complexity balance. Weights adapt to which factors vary most in your data.
| Feature | Free | Standard |
|---|---|---|
| Effort band analysis | Included | Included |
| What-If Simulator (forecast + band match) | Included | Included |
| Story Points / Cycle Time toggle | Included | Included |
| Sensitivity slider | Included | Included |
| Search, sort, compare epics | Included | Included |
| Interactive guide | Included | Included |
| AI Insights (Health Score, Findings, Recommendations) | — | Included |
| Similar Past Epics in Simulator | — | Included |
| AI Prediction (live epic tracking) | — | Included |
| Save and load analyses | — | Included |
| Trend tracking over time | — | Included |
| Scheduled analysis | — | Included |
Standard tier: $2/user/month with a 30-day free trial.
Start with: issuetype = Epic AND status = Done AND resolved >= -6m ORDER BY resolved DESC
This returns all completed epics from the last 6 months. Relativity works best with 20–500 completed epics.
Relativity only analyzes completed epics (Done/Closed status) that have child issues with story points. Epics without completed children or without story points are excluded with a warning.
Story points are summed from completed child issues (issues with a Done status category). The story point field is auto-detected from your Jira configuration.
Currently Relativity only supports Epics. If your JQL returns non-Epic issues, they will be excluded with a warning message.
No. All analysis runs within your Atlassian Forge environment. No data leaves your Jira instance. Relativity has read-only access and cannot modify your Jira data.
Relativity uses proprietary AI-powered pattern recognition to detect natural groupings in your delivery data. The analysis adapts to your specific dataset — weights, thresholds, and band boundaries are all computed from your actual delivery history, not predetermined rules.
© 2026 Relativity Labs. All rights reserved.