Office Conversion Structural Feasibility Multi-Agent with AutoGen and Portia AI
Converting legacy office buildings into residential units hits severe architectural roadblocks such as floorplate depth and plumbing stack limits. Build a multi-agent architectural analysis system using AutoGen and Portia AI to evaluate floor plans and estimate conversion costs per square foot. Achieve less than 8% error in structural feasibility estimates.
What you are building
The core problem, expected build, and operating context for this challenge.
Deploy AutoGen multi-agent workflows integrated with Portia AI automation to assess office-to-residential floorplate conversions.
How work is evaluated
Evaluates structural conversion cost estimation accuracy and feasibility classification.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
feasibility_match
Checks correct classification of conversion feasibility.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
cost_mape
Mean Absolute Percentage Error for estimated cost per square foot. • target: 0.08 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Configure AutoGen multi-agent dialogue loops between Architectural and Structural engineering agents
Utilize Portia AI to execute automated plan analysis workflows and CAD metric extraction
Calculate key conversion feasibility indicators including window distance ratios and plumbing drop capacity
Produce detailed per-sqft cost estimates for adaptive reuse engineering
Reference links and supporting material
Dataset of commercial office floorplate metrics, window ratios, riser counts, and historical retrofit actual costs.
Challenge setup
Evaluates structural conversion cost estimation accuracy and feasibility classification.
Challenge input
JSON object with floorplate_depth_ft, total_sqft, window_perimeter_ft, existing_risers
AutoGen
Assigned mandatory framework for multi-agent negotiation.
Portia AI
Assigned mandatory tool integration.
Evaluated output
JSON object with feasibility_status ('FEASIBLE'|'UNFEASIBLE'), est_cost_per_sqft
- Checks correct classification of conversion feasibility.
- Mean Absolute Percentage Error for estimated cost per square foot. • target: 0.08 • range: 0-1
- Benchmark: ConstructionRetrofitBench
- Cost Mape target: 0.08
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Portia · Required
- AutoGen · Optional
- Helicone · Optional
Evaluation contract
- feasibility_match · Weight 1
- cost_mape · Weight 1
Metrics
- cost_mape · Target 0.08
Benchmarks
- ConstructionRetrofitBench
- materials
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON object with floorplate_depth_ft, total_sqft, window_perimeter_ft, existing_risers
Output: JSON object with feasibility_status ('FEASIBLE'|'UNFEASIBLE'), est_cost_per_sqft
Execution harness: Python
Evaluation tools
- AutoGen · Multi-agent Conversation Orchestrator
- Portia AI · Workflow & Tool Automation Execution
Recipe freeze marker
No recipe freeze time is recorded.