Hi MoT community! 
I’m a graduate researcher at National Taiwan University of
Science and Technology, conducting a study on
AI-assisted BDD documentation generation.
What I’m researching:
I’m investigating how different design inputs (boundary objects
such as HMW statements, wireframes, and business rules) affect
the quality of BDD specifications generated by LLMs, using
Oliveira et al. (2019)'s BDD quality framework as the
evaluation standard.
What I need from you:
I’m looking for 5 professionals with BDD/Gherkin practical
experience to spend approximately 20 minutes evaluating
two AI-generated BDD specifications using a 12-question
quality rubric.
Survey link:
What you get:
- Named acknowledgment in the published thesis (optional)
- Full research findings shared upon completion
The two BDD specifications cover a media platform feature
(newsletter subscription + content browsing). They were
generated using different combinations of design inputs,
so the quality difference should be quite apparent to
experienced practitioners.
Your expert perspective would be invaluable — especially
in highlighting where AI evaluation and human expert
judgment diverge.
Thank you so much! 
Wu Wan-Yu
Graduate Institute of Design, NTUST