We still ship bugs to prod, and AI hasn't fixed that.

Coding agents write mobile features fast, but they can't reliably prove the app works on a real device, and every platform speaks its own automation dialect: XCUITest, Espresso, Appium, Detox.

This talk shows another way: one Playwright-style API that behaves identically across iOS and Android, on real devices, simulators, and emulators. Live on stage, we connect an open-source CLI and framework to a coding agent, let it explore a running app through the accessibility tree.
no vision model, no brittle coordinates catch a real regression, fix it, and re-run to prove the fix. That self-remediation loop wires straight into CI and runs at scale.

You'll leave knowing how to give your agent deterministic device control, standardize test authoring across every mobile stack, and turn agent exploration into tests that guard every commit. All open source.