What's new at AgileAI Labs.

AgileAI Labs is a sponsor of Insurity’s Excellence in AI & Insurance 2026, the P&C conference for carriers and MGAs running Insurity’s core platforms. On Tuesday, October 6, 9:30–10:15 am, Kurt Bedocs, VP, Quality, Insurity, and Missy Trumpler, CEO, AgileAI Labs, present “How to Put Guardrails Around AI-Driven Testing, Quality Agents, and Self-Healing Automation”: where AI-driven testing creates real value, where guardrails matter most, and what to look for as AI-driven quality becomes embedded in insurance technology platforms.

The session’s guardrails are the ones Spec2RunAI, the test executor inside Spec2TestAI, now ships: Assertion Lock, so healing can change how a step runs but never what it verifies; verification coverage on every Run Evidence Report, so a passing run cannot hide unchecked steps; and an AI provenance appendix a reviewer can read. Spec2RunAI has been tuned against a live Insurity environment in partnership with Insurity — see the Insurity test automation page.

Two customer-facing capabilities shipped to Spec2RunAI, the test executor inside Spec2TestAI, on September 27, 2026. Run Evidence Reports give every finished run a printable record of what the engine did at each step, what it checked, and what the screen showed — written for the people who did not watch the run. The headline number is verification coverage: each step is marked Verified, Not verified, Failed, or Not run, with the reason, so a passing run cannot hide unchecked steps. An AI provenance appendix proves the AI never changed what a test verifies and records which version of each test case ran.

Real-browser testing runs the same tests, unchanged, in real Google Chrome or Microsoft Edge on Windows 11, Windows 10, or macOS through the customer’s own BrowserStack Automate account, with the platform, version, and screen resolution chosen per run. See both on the Spec2RunAI page →

AgileAI Labs has made Spec2TestAI's no-code, natural-language automated test executor available. The new capability lets people build and run automated tests by writing in plain English — removing the programming barrier that has long kept business analysts, product managers, and manual testers out of test automation.

It works alongside Spec2TestAI's requirements enhancement, mathematical test generation, and predictive analysis to extend Quality Intelligence across the entire SDLC — and integrates GenRocket's Data Connect™ to provision synthetic test data for sensitive applications at scale.