HomeHow It Works
Automation Workflow

From scanned AWB to auto-filled forms in under 2 seconds

See how KabyTech's KabyTech-OCR engine receives Air Waybill documents in any format, extracts every field with cognitive intelligence, validates against IATA & Thai Customs databases, and pushes structured data into your systems.

End-to-End Process

The complete AWB automation workflow

From the moment an AWB arrives — by email, scan, upload, or API — to structured data in your TMS, ERP, or Thai Customs e-Filing system.

1
📄

Document Input

PDF, scan, photo, email, or EDI text

2
🔍

KabyTech-OCR Read

Cognitive reading — no templates needed

3
🧠

Extract & Score

170+ fields with confidence scores

4

Validate

IATA codes, HS codes, Thai Customs

5
📤

Output & Fill

JSON → TMS / ERP / e-Customs

EMIRATES — AIR WAYBILL
176-1551 5964
────────────────────────
Shipper: KARGO GLOBAL LOGISTICS
CO.,LTD. Bangkok 10260
────────────────────────
Consignee: EXCEL MARITIME
& LOGISTICS PVT LTD
Chennai 600 001 India
────────────────────────
Airport: BKK → DXB → MAA
Flight: EK0375/10SEP
Pcs: 5 Wt: 57.72 KG
────────────────────────
AIRBAG INFLATOR
UN3268 / CLASS 9
FREIGHT PREPAID
Step 1

Accept AWBs from any source

No pre-sorting required. KabyTech auto-classifies MAWB vs HAWB vs Manifest and begins processing immediately — regardless of format, language, or layout.

📄 PDF (scanned or digital) 📷 Camera photos 📧 Email attachments 📡 EDI / Cargo-IMP text 📁 SFTP folder watch 🔌 REST API upload
KabyTech-OCR Engine Processing
📖
Read
OCR reads every character — Thai & English
🧠
Analyze
Structure detection, pattern recognition
🎯
Select
Best candidate by confidence score
PRE-TRAINED ON
500K+
real AWB documents
Step 2

Cognitive KabyTech reading — not templates

Unlike traditional OCR that needs a template per document layout, KabyTech uses cognitive extraction. KabyTech understands document structure, identifies candidate values, and selects the best match — working on any AWB from any airline.

🚫 No template creation 🇹🇭 Thai + English native 📊 Table analysis 📑 Multi-page support ✍️ Handwriting recognition 🔄 Continuous learning
📝 Auto-Filled Shipment Form
AWB Number176-1551 5964
AirlineEMIRATES (EK)
ShipperKARGO GLOBAL
ConsigneeEXCEL MARITIME
Origin → DestBKK → MAA
FlightEK0375/10SEP
Pieces / Weight5 / 57.72 KG
GoodsAIRBAG INFLATOR
DG ClassUN3268 / CL9
ChargesPREPAID
✓ 18 fields extracted · 97.8% confidence
Step 3

Extract every field, fill every form

All 29 FWB sections (170+ fields) extracted with per-field confidence scores. High-confidence fields auto-fill your shipment forms. Low-confidence fields get flagged for one-click human review.

📋 29 FWB sections · 170+ fields 📊 Per-field confidence score ✅ Auto-approve ≥95% 👁️ Human review for <95% 🔄 Corrections improve accuracy 🎯 Fine-tuned per client
Validation Results
✈️Airline prefix EK (Emirates)VALID
🛫Airport BKK (Suvarnabhumi)VALID
🛬Airport MAA (Chennai)VALID
⚠️DG Class 9 — UN3268DG CHECK
🏛️HS Code (Thai Customs Department)VALID
3-Way Matching
📄AWB ↔ ManifestMATCH
📦Manifest ↔ D/OMATCH
💰Charges ↔ Rate CardMATCH
Step 4

Validate everything before it hits your system

Every extracted value is cross-checked against reference databases. Errors are caught before they reach your TMS — not after a Thai Customs rejection.

✈️ IATA airline prefixes 🛫 Airport codes (IATA+ICAO) 🏛️ Thai Customs HS codes 🔢 Tax ID verification 📊 3-way document matching 💰 Charge reconciliation
{
  "awb": "176-1551 5964",
  "airline": "EK",
  "shipper": {
    "name": "KARGO GLOBAL...",
    "country": "TH"
  },
  "routing": ["BKK","DXB","MAA"],
  "pieces": 5,
  "weight_kg": 57.72,
  "goods": "AIRBAG INFLATOR",
  "dg_class": "9",
  "un_number": "3268",
  "charges": "PREPAID",
  "confidence": 0.978
}
Step 5

Structured data wherever you need it

Clean JSON output via API or webhook. Auto-push to CargoWise, SAP, or any TMS. One-click Thai Customs e-filing. No manual re-keying, no copy-paste.

📄 JSON via REST API 📊 Excel / CSV export 🔗 CargoWise connector 🏢 SAP TM / Oracle TMS 🏛️ Thai e-Customs EDI 📁 SFTP batch output
Confidence Scoring

Humans only touch what KabyTech is uncertain about

85%
auto-approved

How it works

≥95% confidence → Auto-approved
~85% of fields flow through with zero human touch
<95% confidence → Human review
~15% routed to review queue — one-click correct, one-click approve
Corrections → accuracy improves
Every correction feeds back into your model — accuracy improves over time
<2s
Processing per AWB
97%+
Field-level accuracy
85%
Auto-approved fields
80%
Time saved vs manual
Before & After

Manual process vs KabyTech automation

❌ Before: Manual AWB Processing

  • 😤 Agent manually reads each AWB document
  • ⌨️ Re-keys shipper, consignee, routing, weight into TMS
  • 📋 Same data entered again into Thai Customs portal
  • ❌ 3–4 Customs rejections/week from keying errors
  • 📈 Cannot scale — more AWBs = more people
6+ hrs
per agent per day
spent on AWB data entry

✅ After: KabyTech Automation

  • 🤖 KabyTech reads any AWB automatically — no templates
  • 📤 JSON pushed directly to TMS / ERP
  • 🏛️ Thai Customs auto-filled from AWB data
  • ✅ Zero rejections — codes validated before filing
  • 📈 Scales automatically — 100 or 10,000 AWBs/day
87%
reduction in data entry time
45 min → 4 min per declaration
Real-World Workflows

Common automation use cases

Ready to automate your AWB workflow?

Book a live demo using your actual AWB documents. See the full workflow — from scan to form fill — in real time.