Skip to main content
Back to challenges
Challenge

Cross-Border Entity Resolution Triage using AutoGen and Kiln AI

Inconsistent corporate entity data across international payment networks causes high false-positive rates in sanctions and AML queues. You will build a multi-agent triage system that reconciles conflicting LEI profiles and customer records, achieving over 95 percent resolution accuracy across ambiguous global payment feeds.

CybersecurityHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Construct an AutoGen multi-agent system tuned with Kiln AI dataset orchestration to automate corporate entity resolution and reduce investigator review backlogs.

Delivery guide

How work is evaluated

Evaluation

Evaluates accuracy of agent consensus in resolving matching corporate identities from ambiguous inputs.

Datasets

Challenge datasets

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

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

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

correct_match_decision

Checks if match determination matches ground truth

Binary check

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

Dimension 2

f1_score

F1 score of entity match decisions • target: 0.96 • range: 0.9-1

Binary check

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

Learning goals

What you should walk away with

  • Orchestrate conversational AI agents in AutoGen for entity matching

  • Integrate Kiln AI for structured dataset generation and prompt tuning

  • Reconcile noisy cross-border merchant LEI data against registry records

  • Measure and lower false-positive alert creation rates in compliance queues

Resources and assets

Reference links and supporting material

Dataset notes

Dataset of 500 cross-border corporate entity record pairs with varying degrees of name, address, and jurisdiction obfuscation.

Challenge setup

Evaluates accuracy of agent consensus in resolving matching corporate identities from ambiguous inputs.

Challenge input

JSON containing entity_a and entity_b records

AutoGen

Multi-agent conversational framework for collaborative consensus building.

Kiln AI

Prompt engineering and fine-tuning suite for structured model outputs.

Evaluated output

JSON with match status ('MATCH'|'NO_MATCH'), confidence, and matched LEI

Evaluation checks
  • Checks if match determination matches ground truth
  • F1 score of entity match decisions • target: 0.96 • range: 0.9-1
Configured evaluation evidence
  • Benchmark: EntityRes-Bench
  • F1 Score target: 0.96
  • 1 public reference case
Configured runtime
  • 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

Action Space
  • AutoGen · Required
  • Kiln AI · Optional
  • Microsoft AutoGen · Optional

Evaluation contract

  • correct_match_decision · Weight 1
  • f1_score · Weight 1

Metrics

  • f1_score · Target 0.96

Benchmarks

  • EntityRes-Bench
  • general

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing entity_a and entity_b records

    Output: JSON with match status ('MATCH'|'NO_MATCH'), confidence, and matched LEI

    Execution harness: Python

Evaluation tools

  • AutoGen · agent_framework
  • Kiln AI · agent_system

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Cross-Border Entity Resolution Triage using AutoGen and Kiln AI