JSON-AI Feedback

Ron Itelman
2 min readNov 2, 2018


One problem I’ve noticed companies have is that various teams across various products will use entirely different user feedback and user ratings systems. This creates an entropic information graph. JSON-AI addresses this problem by having one singular set of numbers for all feedback so that all products and experiences are using the same system. JSON-AI’s Hero Feedback System @i.:) is a set scale from 1–5 that has multiple use cases described below.


  • *.@i.feedback.:) is the same as writing @i.:)
  • This is because the ‘*’ character means that this pattern is reserved as part of JSON-AI’s constellation.



Emotions are a rich and powerful language, and are a standardized component of the JSON-AI schema:

  • Labels to a knowledge snippet’s emotional experience:
    “Many students face daily hunger in America, school is the one meal a day they get. Schools are getting their meal budgets slashed and we don’t know what the impact will be to society.”
  • Knowledge about a user experience:
    “We use JSON-AI for UX Research Reports and have a standardized way to record user feedback.”
@i.:D - Wondrous
@i.:) - Good
@i.:| - Ambivalent
@i.:( - Bad
@i.x( - Terrible

Like / Dislike

@i.:) - like / thumbs up
@i.:( - dislike / thumbs down

Helpful / Not Helpful

@i.:) - Helpful
@i.:( - Not helpful

Star Rating System

@i.:D  - 5 out of 5
@i.:) - 4 out of 5
@i.:| - 3 out of 5
@i.:( - 2 out of 5
@i.x( - 1 out of 5



Ron Itelman

I like human + machine learning systems | Principal — Digital, Data & Analytics at Clarkston Consulting