The quality bar: what gets a trainer suspended, what gets them celebrated

Three ways to see it

1. 1

   AIzaadi publishes its quality bar openly so trainers can self-monitor and so institutions can hold us accountable. Three quantitative SLAs apply to every cohort. Average learner satisfaction must remain at or above 4.0 out of 5 across the trailing four cohorts. Post-cohort capability score, measured by the four-axis assessment from lesson five, must average at or above 65%. Cohort completion rate, the percentage of enrolled learners who finish all twenty-one hours and submit the capstone, must remain at or above 75%. Falling below any one for two consecutive cohorts triggers a coaching review. Falling below any one for four consecutive cohorts triggers suspension.

2. 2

   Way one to stay above the bar: instrument every cohort with continuous feedback, not end-of-cohort surveys alone. A two-question pulse at the end of each three-hour session, satisfaction now and one thing to change tomorrow, gives you fourteen data points instead of one. You can correct course inside the same cohort. Trainers who do this average half a satisfaction point higher than those who only collect end-of-cohort feedback. The instrumentation is not a quality control tax, it is a quality production tool.

3. 3

   Way two: behavioural standards are absolute, not graded. Some failures end the certificate immediately, regardless of satisfaction scores. These include: any form of harassment, sexual or otherwise, toward a learner; any racial, sectarian, or ethnic slur, even in joking; any unauthorised collection of learner personal data; any monetisation of the cohort outside the AIzaadi platform; any plagiarism of curriculum content for personal-brand publication. There is no warning ladder for these. One verified incident triggers immediate decertification. The certificate's value rests on this absoluteness. A platform that punishes harassment in some cases and not others is not a platform Pakistani women, minorities, and PWD learners can trust.

The why-tree

Try this with Claude

AI-edge prompt: 'I am a certified AIzaadi trainer who just delivered three cohorts. My satisfaction scores were 4.6, 4.1, and 3.8. My capability scores were 72%, 68%, and 60%. My completion rates were 84%, 80%, and 70%. Diagnose what is going wrong using only the trend, propose three concrete corrections to my next cohort design, and write me an honest self-review email I can send to my AIzaadi mentor before they have to call me.'