The Test Tribe New York | Testers and QA Meetups

10th New York Meetup – Repairing Playwright Tests with Specialised AI Agents and Graph RAG

When a Playwright test fails, the failure isn't always a broken application.
It could be an outdated locator, a synchronization issue, or a genuine regression. Before repairing the test, you need to understand what actually went wrong.

At The Test Tribe’s 10th New York Meetup, Srinivas Rao Jonnakuti will explore how Graph Retrieval-Augmented Generation (Graph RAG) can connect requirements, source code, tests, and execution evidence to help AI agents diagnose and repair Playwright failures more intelligently.

Event Details

• Date: 28 October 2026
• Time: 06:00 PM – 08:00 PM ET
• Venue: 307 W 38th St #1401, New York, NY 10018, USA

Session – Repairing Playwright Tests with Specialised AI Agents and Graph RAG

Speaker: Srinivas Rao Jonnakuti

Using an illustrative checkout scenario, the session explores a workflow where specialised agents handle different stages of the testing process from planning and test generation to execution, diagnosis, and repair.

You’ll explore:

• How Graph RAG connects related software artefacts to provide context for test diagnosis and repair.
• How specialised agents can divide responsibilities across the testing workflow.
• How traces, network evidence, and repository changes can help distinguish test defects, application regressions, and synchronization issues.
• How AI-assisted repairs can be validated while preserving existing assertions and keeping application defects visible.
• Safeguards such as regression checks, repair limits, and human review for high-risk changes.

How This Meetup Solves Your Problem

• Playwright failures take too long to diagnose? Explore how connected software context can help agents identify the likely cause.
• AI repairs risk hiding real application bugs? Learn how to preserve assertions and validate targeted fixes.
• Not every failure is a test defect? See how execution evidence can help distinguish broken tests from genuine regressions.
• Concerned about autonomous test repair? Explore practical safeguards and where human review remains important.

Why Show Up on 28 October?

• Because fixing a failed test starts with understanding why it failed.
• Because AI-assisted repair is more useful when agents have context across requirements, code, tests, and execution evidence.
• Because specialised agents can divide complex testing workflows into focused responsibilities.
• Because autonomous repair needs safeguards — not just automation.
• Because the goal isn't to make AI fix more tests. It's to make sure it fixes the right tests without hiding real defects.

Become a Speaker at The Test Tribe Events

AI can turn a plain-English test idea into executable mobile automation. The harder part is making sure the generated test follows team standards, remains reviewable, and still leaves the important decisions with the tester.

At The Test Tribe’s 7th North America Huddle, Srikanth Srinivas will demonstrate an AI-assisted mobile automation workflow using WebdriverIO, Appium, and a real Android device.

The second session looks at a different side of quality engineering. Vinitha Kalappurakkal will share how treating testability as an architectural problem helped reduce a retroactive billing process in a healthcare platform from approximately 72 hours to 3.5 hours while maintaining its quality and financial validation.

## Session 1

Prompt to Passing Test: AI-CLI-Driven Mobile Automation
Speaker: Srikanth Srinivas
Srikanth will demonstrate how an AI command-line assistant can take a plain-English testing idea and turn it into an asserting mobile test using WebdriverIO and Appium, then execute it on a real Android device.

## Session 2