Skip to content

Personalization Decision Tree

When a chatbot receives a query, it must decide how much it can and should personalize the answer. That decision depends on whether the user is known, what they are allowed to see, what they have asked for before, and what they prefer. This decision tree walks through that logic from query to delivery.

Interactive Demo

Run MicroSim Fullscreen

You can embed this MicroSim in your own page with the following iframe:

1
<iframe src="main.html" width="100%" height="802" scrolling="no"></iframe>

Overview

The tree branches on three decision points (shown in orange):

  1. Authenticated? — If not, return a generic public-only response.
  2. Has data permissions? — A security-critical check. If not, explain the restriction and offer to request access.
  3. Has preference settings? — Apply explicit preferences if present, otherwise fall back to sensible defaults.

After loading context and analyzing history, a final pattern match decides between an enhanced response (data plus proactive follow-ups) and a standard personalized response. Every interaction is then logged so future personalization improves.

The side panel lists the four personalization data sources (profile, preferences, history, session) and a concrete example user, Bob Martinez, a regional sales manager whose history reveals a recurring quarter-over-quarter comparison pattern. Hover any node to read exactly what the system does there.

Lesson Plan

  • Trace: Have students follow Bob Martinez's query "Show me sales data" down the tree and predict which leaf node it reaches.
  • Discuss: Why is the permission check placed before the preference check? What security problem would result from reversing them?
  • Analyze: Identify which branches depend on explicit data versus learned data, and discuss which is more reliable.
  • Apply: Ask students to add a branch for a brand-new authenticated user with no history yet.

References