LLM OneShot

Last updated: March 2026

1 Concept

LLM OneShot generates a single, personalized message using an LLM.

  • No multi-turn conversation
  • Fully embedded into the rule-based flow
  • App users experience it similar to a normal message, with an AI indicator

The main benefit is for coaching designers:
Highly personalized messages with many variations can be created with minimal effort.

2 Typical Use Cases

  • Personalized feedback messages
  • Motivational messages (e.g. encouraging questionnaire completion)
  • Emotion-aware responses with positive reinforcement

3 Creating an LLM OneShot Message

  1. Go to Micro Dialogs
  2. Add a new message
  3. Select LLM OneShot as message type

Compared to a normal message, two fields differ:

Prompt

Defines what the LLM should generate.

Example:
Greet the user.

Local LLM Meta Information (Optional)

Additional, message-specific meta information.

  • Supplements the global LLM settings
  • Optional

Test Run

Use Test Run in the editor to:

  • Preview an example output
  • See how the LLM interpreted the task

4 Visual Indicators

  • LLM OneShot messages are marked as AI generated
  • OneShots can:
    • have no user reply
    • request free-text input
    • present selectable options

| | |