Why the Packaging Industry Has a Unique Problem
Packaging companies process thousands of orders per year -- via email, as PDF, Excel or free text. Unlike industries with high EDI adoption, orders in the packaging industry are mostly unstructured. The result: enormous manual effort in order entry that scales linearly with order volume.
Heterogeneous Order Formats Without Standardization
- Customers order in completely different formats: PDF orders, Excel spreadsheets, email free text
- No EDI connection for a large portion of customers -- especially in the mid-market segment
- Each customer uses their own item numbers, layouts and sometimes different languages
- EDI often covers only a fraction of order volume -- the rest is manual
Complex Item Matching
- Customers use their own material numbers that need to be mapped to internal ERP item numbers
- Master data reconciliation is labor-intensive: customer items, prices, delivery addresses and terms must be validated
- Distinction between reorders (known item) and new orders (not yet in the system)
- Plant determination: the same item can be produced at multiple manufacturing sites -- the assignment is often rule-based
Growing Volume with Unchanged Resources
- Packaging companies grow -- organically, through new customers or acquisitions -- but the back office doesn't scale accordingly
- Personnel effort scales linearly: twice as many orders means twice as many hours of data entry
- Seasonal peaks (Christmas season, promotions, product launches) create order spikes that the existing team can barely handle
- Temporary staff for peak times is expensive and error-prone -- onboarding takes weeks
High Time Investment per Order
- 3-5 minutes per order on average: open email, review attachment, identify relevant data, search and match in ERP, manually create order
- For a typical mid-sized packaging company with 1,000-2,000 orders per month, this adds up to 600-2,000 hours per year -- pure data entry work
- Qualified staff spend most of their time on repetitive data entry instead of value-adding tasks
The Typical Order Process -- and Where It Fails
The manual order process in the packaging industry follows a typical pattern. At every step, error sources and time wasters lurk.
At every step, friction arises: emails must be filtered, different formats interpreted, customer material numbers manually translated and master data validated. A single error in order creation -- wrong quantity, wrong plant, wrong item number -- triggers a domino effect in production.
Three Order Types -- and Why Not All Are Equally Automatable
In the packaging industry, orders can typically be divided into three categories. This distinction is crucial for a realistic automation strategy.
Standard Orders (Reorders)
- Largest share: ~60-80% of total order volume
- Item is known, customer is known, terms are established -- only quantity and delivery date change
- Fully automatable -- order can be created directly in the ERP, functionally equivalent to an EDI order
- This is where the greatest ROI leverage lies: thousands of orders per year with zero manual intervention
Orders with Deviations
- Known item, but with changes: different delivery address, special terms, deviating packaging units, modified specifications
- The system detects the deviation and routes the order to the responsible clerk for validation
- Partially automated: data extraction and routing run automatically, only the decision remains with the human
First-Time Orders / New Items
- Item does not yet exist in the ERP -- e.g., a new packaging format, a new customer or a completely new product line
- The system automatically detects this and routes the case to the manual process
- Still valuable: automatic data extraction saves significant time even in the creation of new items
The Solution -- AI-Based Order Automation
A modern approach to order automation in the packaging industry covers five phases -- from email to ERP order.
- Document Intake -- Automatic email retrieval, attachment extraction, AI classification (order vs. inquiry vs. complaint vs. not relevant)
- AI Document Processing -- Intelligent extraction of order data (item number, quantity, delivery date, delivery address) from any format -- PDF, Excel, free text
- Business Logic & Master Data Matching -- Map customer material number to internal ERP number, plant determination, address validation, price verification
- Workflow Orchestration -- Automatic decision: process fully automatically (standard order), route for validation (order with deviation) or process manually (first-time order)
- ERP Integration -- Automatic order creation in the ERP (e.g., SAP, Microsoft Dynamics, Oracle), functionally identical to an EDI order
The key advantage: the system automatically distinguishes between the three order types and selects the optimal processing path. Standard orders are created fully automatically -- without any manual intervention.
Savings Potential -- The Numbers Speak for Themselves
The savings depend on order volume, processing time and wage levels. The following model calculation shows which magnitudes are typical.
| Metric | Before (manual) | After (80% STP) |
|---|---|---|
| Orders/year | 18,000 | 18,000 |
| Effort per order | 4 min (manual) | 80% automatic, 20% at 1 min each |
| Personnel effort/year | 1,200 hours | 60 hours |
| Fully loaded cost/year | ~EUR 60,000 | ~EUR 3,000 |
| Savings | ~EUR 57,000/year (95%) | |
- Orders/year
- 18,000
- Before: personnel effort
- 1,200 hours (4 min per order)
- After: personnel effort
- 60 hours (80% automatic, remainder at 1 min each)
- Savings
- ~EUR 57,000/year (95%)
The specific figures vary by company. Use the interactive ROI calculator below for an individual calculation.
Real-World Example: Mid-Sized Packaging Manufacturer
A European packaging manufacturer with multiple production sites processes around 1,500 orders per month. About 60% come via email (PDF, Excel), the rest via EDI. The email orders were previously all processed manually -- by a team of 3 clerks operating at full capacity.
- Starting point: ~10,800 manual orders/year, 4 min processing time, error rate ~3%
- Solution: AI-based order automation with ERP integration and automatic master data matching
- Result after 6 months: 78% STP rate for standard orders, processing time reduced to under 1 minute for orders requiring validation
- Impact: One clerk was reassigned to higher-value tasks (customer service, complaints management)
ROI Calculator for the Packaging Industry
Adjust the values to match your company. The calculation includes fully loaded costs incl. employer social contributions and overhead. For automated orders, only brief validation times apply.
Human-in-the-Loop -- Why 100% Is Not the Goal
A common misconception: automation does not mean completely replacing humans. The "Human-in-the-Loop" (HITL) concept is a deliberate design decision.
- Humans only intervene for exceptions: new layouts, missing master data, changed items
- The system learns from every manual correction -- the automation rate increases continuously
- Validation interfaces show pre-extracted data -- the clerk only confirms instead of retyping
- Error rate drops drastically: machine extraction with human oversight is more accurate than purely manual entry
The goal is not 100% automation, but 100% control with minimal effort.
Why EDI Alone Is Not Enough
EDI (Electronic Data Interchange) is the gold standard for structured data exchange. But in practice, EDI has clear limitations:
- Limited coverage: Only a portion of customers -- mainly large accounts -- have implemented EDI
- Mid-market without EDI: Many packaging customers, particularly in the mid-market segment, still order via email and PDF
- Email remains channel no. 1: In the packaging industry, 50-70% of all orders arrive via email
- EDI implementation is expensive: CHF/EUR 10,000-50,000 per customer for setup and mapping
AI-based order automation bridges the gap between EDI and manual entry. It processes unstructured orders just as reliably as structured EDI messages -- without requiring the customer to change anything.
Learn more: Use Case: Packaging Industry