Automating Agentic AI Success Using This SECRET Workflow
This title could be clearer and more informative.Try out Clickbait Shieldfor free (5 uses left this month).
Dave Farley interviews two practitioners — Stefan Ellisdorfer and Christian Gassel — about using Acceptance Test-Driven Development (ATDD) as a structured workflow for guiding agentic AI coding assistants. The core idea is that well-written, executable acceptance criteria serve as a precise fitness function that AI agents can iterate against, reducing non-determinism and the need for human review of implementation details. Both guests argue that ATDD and BDD naturally align with what agentic AI demands: clear, specific, verifiable specifications of desired behavior. The discussion also touches on the 'Farley index' as a feedback mechanism for scoring AI-generated tests, and the broader shift toward raising the abstraction level of programming through behavioral specifications.