The challenge
Requirements, QA analysis, and automated tests often live in separate tools. Coverage gaps can surface late, while links from business intent to test evidence stay hard to follow.
The workflow
Business context and comments
Acceptance criteria and edge cases
Coverage gaps in Playwright
New tests and Jira results
A closer look
A secure-login story becomes explicit cases for valid access, invalid credentials, and required form controls. The agents compare those cases with the existing suite, create missing tests, and report their results in Jira.
Evidence & status
- Three agents have clearly separated responsibilities.
- Jira keys link stories, bot comments, and Playwright tests.
- Signed webhooks, filtering, deduplication, and per-story queues handle Jira events.
- Separate bot identities make each contribution attributable.
Advanced local proof of concept; not deployed to production. The dashboard uses illustrative data and orchestration state remains in memory. Persistence, monitoring, stronger execution gates, and CI integration are still needed.
An engineering decision
A test that exposes an unfinished interface stays failed. Weakening assertions to make it pass would remove useful evidence.
The opportunity
The opportunity: fewer manual QA handoffs and earlier visibility into requirement and coverage gaps. Production savings, defect reduction, and coverage gains have not been measured.
Adapted from “NAB Oct 2026 — AI Use Cases & Stories.” Statuses and findings reflect that source. Names are listed as project contacts, not credited authors.
Open the project story on LinkedIn