Requirements and AI

Requirements are the starting point for coverage: user stories, PRDs or feature specs. Every test case can link back to the requirement it verifies, so you can see what's covered, what's failing and what isn't tested yet.

Add a requirement

Requirements → Add requirement (or press n on the Requirements page). Give it a title and paste the content. The more specific it is (acceptance criteria, field names, error messages), the better the AI analysis and generated test cases will be.

Import from Jira or Notion

• Jira: connect Jira under Integrations, then use Import from Jira on the Requirements page to choose issues to import. Linked Confluence pages and attachments can come along too. Issues you've already imported are skipped.
• Notion: connect Notion under Integrations, then import pages shared with the integration.

AI analysis

Open a requirement and choose Analyze. The AI reads it and lists:

• Risks: what's most likely to break
• Ambiguities: questions to clarify with the product owner
• Edge cases: unusual inputs and states worth testing
• Coverage gaps: scenarios that no test case covers yet

You can edit any of these. If you choose Regenerate, sqale asks before replacing your edits.

Generate test cases

From a requirement, choose Generate test cases. The AI drafts cases with steps and expected results. Review them, edit as needed, and approve them. Only approved (and stale) cases are included in test runs, and only they count toward coverage and readiness; drafts don't.

Status, due dates and readiness

• Status: Draft → Analyzed → Approved. You can also set it by hand from the requirement's details.

• Due date: set it on the requirement. The Requirements list can filter for overdue items.

• Readiness: the latest results of the requirement's approved test cases. It's the same everywhere in sqale (requirements, traceability, releases, Insights, the agent). The first that applies:

  • No tests: no approved test cases yet
  • Failing: a test failed
  • Blocked: a test is blocked
  • Not run: some tests have no result yet (skipped counts as not run)
  • At risk: every test passed, but a critical or high bug is open on the requirement
  • Ready: every test passed and no critical or high bug is open

  On a release page, readiness uses only results from runs linked to that release (see Plans and runs).

Archive or delete

• Archive hides a requirement from lists, coverage, traceability and search without deleting anything. View archived requirements from the Requirements page and restore them at any time.
• Delete permanently asks what to do with its test cases (keep them unlinked, or delete them too). Delete is blocked when the requirement is part of a released release.

AI credits

AI analysis and generation use a monthly AI budget, shown as credits: 1 credit = $0.01 of AI usage, measured from the tokens each request actually uses (Free: 100 credits, $1; Pro: 1,000 credits, $10). Short analyses use a fraction of a credit's worth; generating a full set of test cases uses more. The counter in the top bar shows what you've used this month. An admin can turn AI features off for the whole workspace under Settings.