Skip to main content

​
Writing Effective Prompts

The quality of your results depends heavily on how you describe what you want. This guide covers principles and techniques for writing effective prompts.

​
Core Principles

​
1. Be Specific

The more detail you provide, the more accurate the result. 
Create a wall on Level 1 from point (0,0) to (10,0),
height 3 meters, wall type "Generic - 200mm"

​
2. Include Context

Tell the AI what you’re working on and why. Context helps it make better decisions. 
I'm designing a residential building.
I need to create a schedule showing all windows
grouped by level, with columns for type, width,
height, and sill height.

​
3. Break Complex Tasks Into Steps

For complex operations, break them into sequential steps rather than one large request. 
Step 1: List all wall types in the project
Wait for response, then:
Step 2: Change all walls on Level 2 that use
"Generic - 200mm" to "Exterior - Brick on CMU"

​
4. Use Correct Terminology

Use proper technical terms that match your software’s API and UI:
  • Revit: “Family type”, “Parameter”, “Level”, “View filter”, “Schedule”
  • AutoCAD: “Layer”, “Block”, “Polyline”, “XData”
  • Rhino: “Surface”, “NURBS”, “Layer”, “Material”

​
5. Specify Names and Values Exactly

When referencing specific elements, use exact names as they appear in your software: 
Set the "Fire Rating" parameter to "2 HR"
for all walls of type "Interior - 138mm Partition (1-hr)"

​
Prompt Patterns

​
Query Pattern

Ask questions about your model: 
How many doors are on Level 1?
What is the total wall length on Level 2?
List all rooms that don't have a department assigned.

​
Create Pattern

Create new elements with specific properties: 
Create a floor plan view for Level 3 named "Level 3 - Furniture"
Place a 900x2100mm single-flush door in the wall between Room 101 and Room 102

​
Modify Pattern

Change existing elements: 
Change the color of all structural columns to red in the active view
Set the "Mark" parameter to "W-{number}" for all windows, numbering sequentially

​
Analyze Pattern

Get insights and summaries: 
Summarize the areas of all rooms grouped by department
Check if any walls are shorter than 500mm — these might be modeling errors

​
Common Mistakes

Avoid these common pitfalls:
  • Being too vague: “Fix the model” — the AI needs specifics
  • Assuming context: The AI can see your model data, but it doesn’t know your project history or intentions unless you tell it
  • Chaining too many operations: Break complex workflows into individual steps
  • Not specifying exact names: “That wall type” vs. the actual type name

​
Application-Specific Guides

For detailed prompting tips tailored to your software:

Revit

Revit-specific tips and examples

AutoCAD

AutoCAD-specific tips and examples

Rhino

Rhino-specific tips and examples