Generative Floor Plans and Site Design

From Program to Plan

🔄 Quick Recall: In the previous lesson, you learned to generate concept renders and visualizations with AI. Now you’ll go deeper — using AI to create the floor plans and site layouts that define how a building actually works.

Traditionally, generating floor plan options is one of the most time-intensive parts of early design. An architect might spend days creating three or four layout variations for a client. AI generative design tools can produce dozens of feasible options in minutes — not replacing your design judgment, but giving you far more material to evaluate and refine.

Generative Floor Plan Tools

Maket.ai — Residential Floor Plans

Maket.ai is the most accessible generative floor plan tool. It works by taking your constraints and producing layout options:

Generate floor plan options using these constraints:

Building type: [residential / commercial / mixed-use]
Total area: [square feet/meters]
Required spaces:
- [Room 1]: [size range], [adjacency requirements]
- [Room 2]: [size range], [adjacency requirements]
- [List all program spaces]

Constraints:
- Lot dimensions: [width × depth]
- Setbacks: [front, side, rear]
- Orientation: [which direction does the lot face?]
- Parking: [number of spaces required]
- Accessibility: [ADA/universal design requirements]

Preferences:
- Open plan vs. defined rooms
- Natural light priority spaces: [which rooms need the most daylight?]
- Main entry approach: [from street, from parking, from garden]

Generate [number] distinct layout options with room labels and area calculations.

TestFit — Multi-Unit and Commercial

TestFit specializes in real estate feasibility — generating configurations with real numbers:

What you input: Site boundary, setbacks, zoning envelope, target unit mix (studio/1BR/2BR/3BR), parking ratio, amenity requirements.

What you get: Multiple building configurations with unit counts, parking layouts, estimated costs per unit, and gross-to-net ratios. Changes are real-time — adjust a parameter and watch the plan update instantly.

Best for: Multifamily residential, mixed-use, and commercial feasibility studies where developers need numbers to make go/no-go decisions.

Finch 3D — Spatial Optimization

Finch 3D works inside your Revit/Rhino environment, optimizing spatial relationships:

What makes it different: Instead of generating layouts from scratch, Finch refines your initial layout to optimize for criteria you select — daylight, views, circulation efficiency, or structural grid alignment.

✅ Quick Check: Why should you still manually review AI-generated floor plans even when they meet all your program requirements? Because AI optimizes for measurable criteria (area, unit count, efficiency) but doesn’t experience the space. A plan that’s technically efficient might have dark hallways, awkward room proportions, or spatial relationships that don’t feel right when you walk through them.

AI-Assisted Site Design

Environmental Analysis with Forma

Autodesk Forma (formerly Spacemaker) brings AI to early-stage site analysis:

Solar analysis: Model sun exposure across your site throughout the year. Identify which facades get direct sunlight, where shadows fall, and how massing choices affect neighboring properties.

Wind comfort: Simulate wind patterns around your proposed massing. Predict outdoor comfort levels for terraces, entries, and public spaces.

Noise mapping: Overlay noise data from roads, transit, and other sources. Position quiet spaces (bedrooms, offices) away from noise and buffer spaces (corridors, storage) toward it.

Daylight analysis: Evaluate interior daylight levels for different massing options. Meet daylight factor requirements without detailed energy modeling.

Site Analysis with ChatGPT/Claude

For projects without specialized software, AI assistants help structure your analysis:

Help me analyze this site for a [project type]:

Location: [address or description]
Lot size: [dimensions]
Zoning: [zoning designation and key requirements]
Adjacent uses: [what's next door on each side]
Topography: [flat, sloping, describe grade changes]
Existing features: [trees, structures, water, utilities]
Solar orientation: [which direction does the main frontage face?]

Analyze:
1. Optimal building placement given solar orientation and setbacks
2. Best entry and circulation approach from [street/parking/transit]
3. Where to position outdoor spaces for maximum comfort
4. Potential challenges (views to block, noise sources, drainage)
5. Opportunities (views to frame, prevailing breezes, mature trees to preserve)

Present as a site analysis summary with recommended design strategies.

The Design Brief Generator

Before generating plans, use AI to sharpen your design brief:

Help me develop a complete architectural program for a [project type]:

Client: [who is the client and what do they need?]
Budget: [if known]
Site: [basic site info]
Key requirements they've stated: [list what the client has asked for]

Generate a detailed program including:
1. Complete space list with area ranges (net and gross)
2. Adjacency diagram — which spaces must be next to each other
3. Circulation requirements — public vs. private, service vs. visitor
4. Environmental requirements per space (daylight, acoustics, temperature)
5. Questions I should ask the client that they probably haven't considered
6. Code implications for this building type and occupancy

Format as a design brief I can present to the client for approval before starting design.

✅ Quick Check: What advantage does AI site analysis provide over traditional manual analysis? Speed and comprehensiveness. AI can model solar exposure, wind patterns, and noise levels simultaneously across multiple massing options in minutes. Manual analysis of the same scope would take days. This means you can evaluate more design options with better environmental data before committing to a direction.

Parametric Design with Grasshopper

For architects using Rhino, Grasshopper’s visual programming enables parametric design that responds to AI-analyzed data:

Workflow: Forma (environmental data) → Grasshopper (parametric model) → Rhino (3D output)

Example: Use solar analysis data to drive facade perforation patterns. More openings on north facades (in the Northern Hemisphere), less on west facades where overheating is a risk. The parametric model adjusts automatically as you change building orientation.

This isn’t beginner territory — but if you’re already using Grasshopper, AI environmental data makes your parametric definitions far more responsive to real site conditions.

Key Takeaways

• Generative floor plan tools (Maket, TestFit, Finch) produce multiple feasible layouts from your constraints in minutes
• AI site analysis (Forma) models solar, wind, noise, and daylight conditions to inform massing decisions with evidence
• Always manually review AI-generated plans for spatial quality, circulation logic, and the human experience of the space
• Structured design briefs and program documents make AI-generated options more useful and relevant
• Parametric design workflows can directly incorporate AI environmental data for responsive design

Up Next: You’ll learn to use AI for sustainable design — energy modeling, material optimization, and the environmental analysis that makes green architecture practical, not just aspirational.