Skip to content
Victorian Government · Asbestos Compliance

AsbestosCompliance,ReinventedwithAI

Open AppRead Docs
0%Extraction Accuracy
0BAR Columns
0Stories Shipped
0sSeconds Per PDF
Infrastructure Status
Last checked: checking…

How It Works

From PDF to Compliance in Three Steps

ACM-AI eliminates weeks of manual data entry. Upload your asbestos register and receive a submission-ready BAR workbook in under 30 seconds.

01

Upload PDF

Drag and drop your existing asbestos register PDF — any format, any structure. The system handles digital and scanned documents alike via automatic PDF classification.

02

AI Extracts Data

A 7-stage agentic pipeline powered by LangGraph, MinerU and Docling parses every table — including merged cells and multi-page registers — mapping 47 BAR fields with 96% accuracy.

03

Export BAR Excel

Download a government-ready BAR-compliant Excel workbook via openpyxl. Every row is traceable back to its source PDF page. Submit directly — no reformatting required.

Under the Hood

7-Stage Extraction Pipeline

Every PDF is routed through a precision agentic workflow. Each stage handles a distinct aspect of the extraction — no black-box guessing.

PREPRE-ANALYSIS

TOC extraction, building inventory compilation, page-level section tagging, document metadata enhancement

LangGraph agentic orchestrator

0PREFLIGHT

PDF classifier detects digital vs scanned. Parser router selects MinerU or Docling based on content type.

PDF Classifier → Parser Router

0.5ORCHESTRATION

Agentic orchestrator routes each page section to optimal extraction tool. MinerU for complex tables, Docling for text/layout.

MinerU (primary) | Docling (fallback)

1EXTRACT

Raw table extraction preserving bounding boxes, page numbers, merged cells, and multi-page stitching.

MineruTableExtractor | Docling | Generic Configurable Parser

2INTERPRET

AI maps raw extracted cells to 47 BAR field schema. Wording normalisation, enum validation, product classification (T1–T8).

LLM + register_row.schema.json + register_enums.json

2.5VALIDATE

Validates enum fields against BAR controlled vocabulary. Up to 3 LLM correction attempts for invalid values.

ValidationIssue → CorrectionStats → LLM re-extraction

3SAVE & INDEX

Deduplication with SHA-256 composite keys. Contextual embeddings for vector search. Persisted to SurrealDB.

SurrealDB | Vector embeddings | Parent Document Retrieval

Data stages
AI / LLM stages
Infrastructure

By the Numbers

Built for Enterprise Scale

Every metric reflects real delivery — no aspirational targets. ACM-AI is feature-complete and ready for Victorian Government deployment.

0
Stories Delivered

Across 12 active sprints

0
Epics Complete

Full feature coverage

0
Commits

Autonomous development velocity

0
BAR Columns

Full Victorian Government schema

0%
Extraction Accuracy

Against manual ground truth

0s
Processing Time

End-to-end per document

Feature Complete— All 16 epics delivered as of 22 Feb 2026
Production Ready

Built for Every Stakeholder

What ACM-AI Means for You

Select your role to see how ACM-AI delivers value across the Victorian Government compliance chain.

Risk Mitigation + Operational Efficiency

Hours saved per document: 8–12 hours of manual data entry eliminated

Accuracy: 90%+ extraction accuracy vs error-prone manual entry

Compliance: 100% BAR field coverage — no submission gaps

Traceability: Every data point linked back to source PDF page

Key Insight

Reduces liability exposure from incomplete asbestos registers

Asset Directorperspective