How are PRs considered qualified in the AI Impact report when using traces?
Last updated: August 28, 2026
This article applies when your AI Impact report is powered by IDE traces (Cursor, Copilot, Claude Code, and similar). Span attributes lines in a pull request to the author’s AI sessions. That only works when we have those traces for the life of the PR.
If your report uses Span-detect-1 instead of traces, see How are PRs considered qualified in the AI Impact report.
Qualified PRs are not “PRs that used AI.” It is the share of merged PRs Span can include in the AI code ratio: PRs where we have enough trace coverage to measure AI vs human lines reliably. AI usage itself is the AI Code Ratio column.
Why some PRs are left out
If traces were missing, started after the PR was opened, or had gaps while the PR was in flight, including the PR would mean guessing at AI usage. Those PRs stay in the merged-PR total and are excluded from the ratio.
The report therefore shows two numbers:
Column | What it answers | How it is calculated |
|---|---|---|
AI Code Ratio | How much of the measurable diff was AI-authored? | AI-attributed lines ÷ eligible lines, only on qualified PRs |
Qualified PRs | How much of the author’s merged work could we measure? | Qualified merged PRs ÷ all merged PRs in the range |
A low qualified-PR percentage does not mean people are not using AI. It usually means many PRs in the range were opened before traces were being collected, or that traces were incomplete for those PRs. Read AI Code Ratio for adoption; read Qualified PRs for measurement coverage.
Which PRs are in the denominator
Every PR in Qualified PRs (the “of N”) is a merged PR in your selected date range whose author is a development contributor, and whose repository, branches, and author are not ignored in Settings → Metrics.
The date range buckets PRs by merge date. Qualification itself looks at when the PR was opened, because that is when we need traces to start. A PR opened in April and merged in July still does not qualify if trace collection started in June.
When a PR is qualified
A merged PR is qualified when at least one of the following is true:
Strong trace match. At least 20% of the PR’s eligible diff is attributed to the author’s traces.
Continuous traces, even if little or no AI was matched. Span has an analysis result for the PR, the author’s latest trace is at or after merge, and there is no run of two or more weekdays without traces (excluding out-of-office days) around the PR — from 3 days before it opened through 2 days after it merged. A 0% AI PR can still qualify.
Fully agent-authored. Every non-merge commit is agent-authored (or the author is a background agent with no person in the loop), and the PR has a usable ratio.
You do not need to have used AI for a PR to qualify. You do need traces that cover the PR.
When a PR is not qualified
To the right of the AI code ratio chart, click the underlined Qualified: X out of Y total PRs to open the breakdown. Reasons map to these groups:
Reason in the report | What it means |
|---|---|
Author not traced at the time | The PR was opened before trace collection started for the organization, or before this author’s first trace. Merge date does not matter. |
Gaps in author’s trace data | There is no person author, the PR merged after the author’s last observed trace, or there is a gap in traces while the PR was in flight. Span cannot confirm the diff can be attributed completely. |
All code excluded by settings | Every changed line is excluded by your ignore rules, so there is no eligible code to analyze. |
Background agent, not traced | The PR involved partial collaboration with a background agent. Trace-based attribution does not apply to those PRs. Fully agent-authored PRs can still qualify (see above). |
PR too large to analyze | The diff exceeds the size limit even after ignore rules, so attribution was not run. |
Session too large to analyze | The only traces overlapping the PR were too large to load into memory, so they were skipped. A trace existed but was excluded for size, so this is not a clean “no AI” result; it is re-checked on later runs in case a smaller overlapping trace appears. |
Too much edit history to analyze | Matching traces contained so much AI-edit history (a file rewritten many times over) that attribution was skipped to avoid an excessively long computation. |
Inconsistent diff data | The PR’s diff accounting is inconsistent (more AI lines than total lines), so a trustworthy ratio cannot be computed. |
Awaiting processing | Attribution has not finished for this PR yet. |
How to inspect the PRs
The page title is AI Impact. Find it under AI Effectiveness → AI insights → Comparison. The Qualified PRs percentages in the people or team table are not links.
Open that page and select the AI code ratio card.
To the right of the chart, click the underlined Qualified: X out of Y total PRs.
In the drawer, open Non-qualified PRs, then View PRs. Each PR is listed with the reason it was excluded (Filter by PR AI Traces Qualified = No).
To see one person or team, filter the report first, then click the same Qualified link. The list follows the current filters.
The date picker cannot start before your organization’s trace-collection start. PRs merged in the selected range can still be unqualified if they were opened before that start.
Bottom line
Measure AI adoption on work Span can actually attribute from traces. Qualified PRs is coverage of that measurement, not a usage score. Once traces have been collecting for a while, the qualification rate typically rises; AI Code Ratio on those PRs is the adoption number to compare across people and teams.