Pigai — China's Large-Scale AI Essay Feedback Platform: Real Reach, Mixed Classroom Evidence

Pigai is a Chinese automated writing evaluation (AWE) platform for English-as-a-foreign-language students, in operation since 2011 and run by Beijing Word Network Technology Co.

To provide rapid automated feedback on English essays across vocabulary, sentence structure, organisation and content, supporting EFL writing instruction at scale.

Essays are scored via a four-part rubric (vocabulary 43%, sentence structure 28%, organisation 22%, content 7%) combining error-detection algorithms and language models, delivering feedback within seconds; the platform is used across thousands of schools and universities.

Pigai reaches an estimated 6,000-13,000+ schools and universities and 13-20 million students, with over 200 million essays scored cumulatively; peer-reviewed research finds reliable grammar-accuracy improvements and increased revision frequency, but feedback is largely limited to sentence-level mechanics and misses meaning-dependent errors, and adding Pigai to peer feedback showed no revision-quality improvement beyond peer feedback alone.

Pigai has achieved substantial reach and demonstrated grammar-level benefits, but independent research finds no added value over peer feedback for revision quality and no evidence on meaning-level writing improvement.