AI + Automation Governance: Orchestrating Independent AI Triage Scripts

Support
3

From what I can see, most of my use cases can be done in Discourse with AI Triage and Automation, if one can trigger then next.

Here is a hypothetical flow where each step would send the post contents and a prompt to the LLM:

  1. Check if spam
  • If spam, then hide and flag
  1. If not spam, determine products discussed
  • Add product label(s)
  1. Next determine intent:
    ** Complaint
    ** Question
    ** Suggestion
    ** Information sharing
    ** Positive feedback
  • Add intent tag
  1. If ‘complaint’ then evaluate for hot button issues: ie. contains hot-button phrases (cancel, terrible, slow)
  • If hot button issue, add ‘hot’ tag and assign to Sam
  1. If the intent is ‘positive feedback’ and product ‘wireless phone plan’, write a custom invitation to referral program and send as PM.
2 Likes