Skip to main content
Loading navigation
Back to challenges

Oil Project Regulatory Compliance Assistant with AI SDK

Kuwait's $381m upstream oil approval process requires continuous cross-checking of environmental impact assessments (EIAs) against Kuwait EPA regulations. Build a web-based conversational assistant using Vercel AI SDK (TypeScript) with streaming tool calling to analyze submission packages. Provide zero-latency streaming audit results with 95% rule match precision.

Business OperationsHosted by Vera
Sign in free to start

What you are building

Develop a TypeScript compliance agent using Vercel AI SDK to stream real-time environmental permit verification checks.

How work is evaluated

Evaluation

Evaluates streaming output response validity and accuracy of regulatory breach identification.

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

non_compliant_item_flagged

Correctly flags flaring rate violation

Binary check

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

Dimension 2

compliance_rule_precision

Precision in identifying genuine regulatory violations • target: 0.95 • 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 TypeScript backend leveraging Vercel AI SDK core functions (`generateText`, `streamText`)

  • Define typed tools using Zod schemas for environmental parameter verification

  • Implement client-side UI streaming using `@ai-sdk/react` hooks

  • Validate oil field environmental impact submittals against KOC and KEPA thresholds

Resources

Dataset notes

20 environmental impact submittal packages for Kuwait onshore oilfield developments.

Challenge setup

Evaluates streaming output response validity and accuracy of regulatory breach identification.

Challenge input

JSON object with proposed water reinjection volume and flaring limits

AI SDK

Provides standard streaming and component bindings for web platforms.

Evaluated output

JSON containing status, non_compliant_items array, and recommended remedy

Evaluation checks
  • Correctly flags flaring rate violation
  • Precision in identifying genuine regulatory violations • target: 0.95 • range: 0-1
Configured evaluation evidence
  • Benchmark: RegCheckBench
  • Compliance Rule Precision target: 0.95
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript 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
  • Vercel AI SDK · Required
  • Synthflow · Optional
  • AutoKeras · Optional
Orchestration
  • Vercel AI SDK · Required

Evaluation contract

  • non_compliant_item_flagged · Weight 1
  • compliance_rule_precision · Weight 1

Metrics

  • compliance_rule_precision · Target 0.95

Benchmarks

  • RegCheckBench
  • security

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON object with proposed water reinjection volume and flaring limits

    Output: JSON containing status, non_compliant_items array, and recommended remedy

    Execution harness: JavaScript

Evaluation tools

  • AI SDK · TypeScript AI Framework

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about Oil Project Regulatory Compliance Assistant with AI SDK