Can AI write test cases from Jira tickets?

Yes. The agent reads the ticket and its acceptance criteria and writes the Playwright test. Two engineers review it in a pull request before it merges.

How do AI agents write Playwright tests?

Through the Playwright MCP server, the agent opens your app in a real browser, follows the selector rules in your repo and writes the test in TypeScript. It does not guess selectors from screenshots.

How does the agent learn our app?

Through rules and skills files in plain markdown that we write for your app. When your requirements change, we update the files. We do not train a model on your data.

Who owns the agents and tests?

You do. Everything lives in your repo and keeps working after our engagement ends. Your team only needs its own API key.

We use Cypress or Selenium. Can you migrate us?

We have built and used a migration agent for Robot Framework. For other frameworks, we assess the effort on the strategy call before promising anything.

Agentic QA

Agentic QA Testing, Reviewed by Engineers

AI agents write and maintain your Playwright tests. Two engineers review and approve every test before it merges, and you own everything we build.

Book a Free QA Strategy Call

View the Fabletics Case Study

48h

Free QA plan

Why This Service

What Agentic QA Means

AI agents do the repetitive work of writing and updating tests. QA engineers decide what to test and approve what ships. We run the agents in Claude Code and GitHub Copilot with the official Playwright MCP server, so they work in a real browser against your real app. No agent output reaches your main branch without a code review.

Agentic QA is one part of our wider AI test automation service.

Our Approach

How It Works

01

We Write the Rules

Selector rules and testing skills live in plain markdown files in your repo, written for your app. You can read every rule the agent follows.

02

The Agent Drafts Tests

It starts from your Jira tickets and their acceptance criteria, a page URL or your existing test cases, and writes Playwright tests in TypeScript.

03

Two Engineers Review Every Test

Each AI-written test goes through a pull request. Two engineers review it and approve it before it merges. Tests that are wrong get fixed or dropped.

04

Your Pipeline Runs Them

Tests run in Jenkins or GitHub Actions when developers merge code, in nightly builds and on demand. When a test fails, an AI step analyzes the failure and posts a summary to Slack.

When your app changes, we update the rules files and the agent writes tests to the new rules.

Example

What a Rules File Looks Like

The agent follows rules your team can read. Here is a simplified example of a selector rules file:

# Selector rules (example)- Use getByRole or getByTestId first.- Never select elements by CSS class or XPath.- After login, wait for the dashboard heading before the next step.- Keep one user journey per test file.

Each client's rules are written for their own app and grow as the app changes.

What's Included

What You Get

01

An Agent Set Up for Your App

Rules and skills files written for your product.

02

Tests From Your Tickets

Playwright tests built from your Jira acceptance criteria.

03

Two Reviews on Every Test

Two engineers approve each test before merge.

04

CI Pipelines

Smoke and regression runs in Jenkins or GitHub Actions.

05

Failure Analysis in Slack

Each failed run comes with an AI summary of the likely cause.

06

Migration From Robot Framework

Our migration agent converts Robot Framework tests to Playwright.

07

You Own Everything

Agents, rules files and tests stay in your repo.

Fit Check

When Agentic QA Fits

Good Fit

Web apps with Jira tickets that carry acceptance criteria.
Teams with an existing Playwright or Robot Framework suite.
Projects with a CI pipeline in Jenkins or GitHub Actions.

Not a Fit Yet

A prototype whose screens change every week.
Tickets with no acceptance criteria.
Teams that want AI-written tests merged without human review.

Case Study

Proof: Fabletics

For Fabletics and Savage X, our team moved every UI test from Robot Framework to Playwright with a migration agent, then used a test generation agent to grow coverage. Today 1,000+ automated tests (API and UI) run in the release pipelines, and one service alone runs 650+ UI tests across 8 Jenkins pipelines in about 20 to 25 minutes.

Read the Fabletics Case Study

Demo

See It Live

On request, we demo both agents on the free strategy call: the migration agent converting a Robot Framework test, and the test generation agent writing a Playwright test from a Jira ticket.

Book a Free QA Strategy Call

Technology Stack