Skip to main content
Loading navigation
Back to challenges

UK Planning Portal Zoning Compliance Review System using AutoGen

Municipal planning departments face massive backlogs evaluating UK building permit submittals against local plan policies. Build an AutoGen multi-agent system to cross-evaluate planning drawings, land restrictions, and heritage constraints with 92% automated validation accuracy.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Construct an AutoGen multi-agent conversational review process simulating dialogue between a Applicant Agent, Planning Officer Agent, and Conservation Specialist Agent.

How work is evaluated

Evaluation

Evaluates planning application recommendation precision across permitted development and full planning permission test suites.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

conservation_area_check

Ensures extensions >3m in conservation areas trigger policy conflict flag

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

decision_accuracy

Accuracy of planning decisions compared to benchmark ground truth • target: 0.92 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Build a multi-agent conversational loop in AutoGen with dedicated role descriptions.

  • Parse UK planning application metadata against Use Classes Order rules.

  • Incorporate automated human-in-the-loop review for borderline heritage zone applications.

  • Achieve 92% evaluation precision compared against human planning decision records.

Resources

Dataset notes

50 UK planning application case files with applicant submittals, conservation area maps, and council approval decisions.

Challenge setup

Evaluates planning application recommendation precision across permitted development and full planning permission test suites.

Challenge input

JSON containing application_type, conservation_area, height_extension_m, existing_use_class

AutoGen

Framework assignment for orchestrating multi-role dialogical reasoning.

Evaluated output

JSON with recommended_status, required_documents array, policy_conflict boolean

Evaluation checks
  • Ensures extensions >3m in conservation areas trigger policy conflict flag
  • Accuracy of planning decisions compared to benchmark ground truth • target: 0.92 • range: 0-1
Configured evaluation evidence
  • Benchmark: AutoGen-Gov-Bench
  • Decision Accuracy target: 0.92
  • 1 public reference case
Configured runtime
  • Python execution harness
  • Python sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • AutoGen · Required
  • RAI · Optional
  • Alibi Detect · Optional

Evaluation contract

  • conservation_area_check · Weight 1
  • decision_accuracy · Weight 1

Metrics

  • decision_accuracy · Target 0.92

Benchmarks

  • AutoGen-Gov-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing application_type, conservation_area, height_extension_m, existing_use_class

    Output: JSON with recommended_status, required_documents array, policy_conflict boolean

    Execution harness: Python

Evaluation tools

  • AutoGen · Multi-Agent Conversational Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about UK Planning Portal Zoning Compliance Review System using AutoGen