IFS.ai Customer Order Automation

The feature is designed for roles involved in sales order entry and customer order processing, including:

• Sales coordinators responsible for registering customer orders from customer Purchase Order documents

• Inside sales staff handling emailed or scanned purchase orders

• Customer service representatives creating orders on behalf of customers

• Order administrators reviewing and correcting scanned order data before release

Use this feature when you need to:

• Create a customer order quickly from a scanned or uploaded purchase order

• Reduce manual typing of customer names, PO numbers, delivery dates, and line details

• Process purchase orders in PDF, image, or common office document formats

• Match extracted customer and delivery address information agains existing master data

By using this feature, you can:

• Have order header and line data extracted and filled in automatically

• Spend less time on manual entry and re-keying

• Review and adjust extracted data, customer matches, and part lines before saving

• Reach the review step even when the cusomer cannot be matched automatically - you can select the cusomer manually.
  

Access this feature from the Scan Customer Order Assistant.

Page: Scan Customer Order

Produc area: Sales Order

Location: Navigator →  Sales Order →  Scan Customer Order

The assistant guides you through these steps:

• Scan Order - upload the document and start extraction

• Review Order - review and correct header fields and part lines

• Order Processing Parameters - set release and confirmation options

• Finish - create the customer order

On the Review Order step, you can view the uploaded document in the image viewer while reviewing extracted fields.

When you upload a document and proceed from the Scan Order step, the feature reads the uploaded file and extracts information to populate fields on the Review Order step.

The feature reads from the document:

• Customer name and tax ID / association number

• Customer purchase order number

• Wanted delivery date

• Shipping / delivery address

Line items: part number, description, unit of measure, and line delivery date

It also uses:

• Customer master data to match the extracted customer name or tax ID to an existing customer

• Customer delivery addresses to match the extracted shipping address to a valid ship-to address for the customer

• Sales part and customer part cross-reference to suggest catalog numbers for extracted part codes and descriptions

• Customer defaults for order type, currency, language, and default delivery address when a customer is identified

When you upload a document and proceed from the Scan Order step, the feature reads the uploaded file and extracts information to populate fields on the Review Order step.

The feature does not use:

• Data from other customer orders or unrelated business objects

• Master data you do not have permission to read

• Information that is not present or legible in the uploaded document

• Shipping / delivery address
  
  If you change the Customer on the review step, the feature re-evaluates the delivery address match using the shipping address extracted from the document and the newly selected customer.

After the document is read, the system tried to match extracted text to existing data in IFS Cloud. Matching runs in a fixed order for header fields; part lines are matched when you reached the review step.

For implementation details, thresholds, and APIs, see Reference Data Matching Logic in the technical documentation.

Customer Matching

The system selects a Customer using the tax ID / association number and customer name from the document.

Step 1 - Association number (tax ID)

• The scanned tax ID is compared to customer association numbers in your company.
• A customer is set only when exactly one customer matches.
• If more than one customer shares the same association number, Customer is left empty for you to choose.

Step 2 - Customer name (if Step 1 did not find a customer)

• The scanned customer name is compared to customer names in your company.
  
• The closest name match is used only when the match is strong enough (high similarity).
  
• If the match is not strong enough, Customer is left empty.

Excluded customers

Customers marked to be excluded from scan order processing are never selected automatically. 

What you see on review 

• The shipping / ship-to address text from the document is compared to delivery addresses registered for that customer in customer order setup.
  
• Only addresses that are valid ship-to locations for customer orders are considered — the same addresses you can pick in Delivery Address on the review screen.
  
• When a close match is found, that address is suggested as Ship Address.
  
• When no close match is found, the customer's default delivery address is used instead.
  
• Changing customer on review If you change Customer, the system compares the same scanned shipping address text against the new customer's delivery addresses and updates Delivery Address accordingly. What you see on review
  

• The system searches for a catalog number in this order:
  
• Exact match on customer part number (using scanned part code)
  
• Exact match on customer part number (using scanned description) Exact match on sales part / catalog number (using scanned part code)
  
• Exact match on sales part (using scanned description)
  
• Fuzzy match on customer parts (when text is similar but not identical)
  
• Fuzzy match on sales parts (description-based)
  
• When a match is found, Catalog No and related line details (description, unit, converted quantities) are filled in.
  
• When no match is found, the line is still created with the scanned part number, description, quantity, and unit so you can select Catalog No manually.

What you should do: Review every line. Correct Catalog No, quantity, and unit of measure where the suggestion does not match the purchase order.
 

Supported Documents

Supported formats for extraction:

PDF
PNG, JPEG, JPG, HEIF, TIFF
XLSX, DOCX

Quality considerations:

• Documents should contain clear, legible text (printed or high-quality digital text)
  
• PDF purchase orders with structured tables generally produce better line extraction
  
• Faded, low-resolution, or damaged scans may produce incomplete or incorrect results
  
• Handwritten purchase orders are not reliably supported
  
• Very complex or multi-page layouts may require manual correction of line items

Unsupported or limited scenarios:

• Documents in unsupported formats are rejected or cannot be processed
  
• If the document quality is too poor for the system to read, you receive a message that the document cannot be processed
  
• For PDFs, line items may sometimes be missing; you can add or correct lines manually on the review step
  

Understanding the Output

The feature populates fields based on the content of the uploaded purchase order. Results are presented on the Review Order step for your review before the customer order is created.
For why Customer, Delivery Address, or Catalog No are filled or left empty, see How Matching Works.

Header fields that may be populated:

Line fields that may be populated:

Not all fields may be populated. Results depend on the quality and content of the uploaded document.
 
Example scenarios:

• Clear printed PDF purchase order: Customer name, PO number, delivery date, shipping address, and multiple part lines are populated. Customer and ship address may be matched automatically if they closely match master data.
  
• Scanned image with moderate quality: Header fields may be populated but customer or ship address matching may be incomplete — you select the correct customer and delivery address manually.
  
• PDF with poor table layout: Header data may be extracted but some or all line items may be missing. You add or correct lines on the review step before creating the order.
  
• Customer not found automatically: Scanned customer name is shown, but Customer is empty. You select the customer from the list; the system then applies customer defaults and re-evaluates the delivery address match.
  
• All populated values are suggestions. Review every field and line before finishing the assistant.
   

Enabling and Using the Feature

The feature is available when IFS AI services are enabled, the Scan Customer Order assistant is available in your environment, and you have access to create customer orders. 

Prerequisites

IFS AI / ML services must be enabled in the environment

• You must have access to the Scan Customer Order navigator entry and related customer order permissions
  
• Customers, sales parts, and delivery address setup must exist in the system for matching to work
  
• A valid Site and Order Coordinator must be selected on the Scan Order step

Configuration of AI models and recipes is described in the technical documentation. No additional end-user configuration is required beyond standard Scan Customer Order access.

Upload and Extract

• Open Sales Order → Scan Customer Order.
  
• On the Scan Order step, select Site and Order Coordinator.
  
• Click Add File and upload a purchase order in a supported format.
  
• Click Next to upload the file and start extraction.
  
• The system processes the document and advances you to the Review Order step when header extraction succeeds.

Review and Refine

• On the Review Order step, review header fields such as Customer, Customer PO No, Wanted Delivery Date, and Delivery Address.
  
• Compare populated values with the document shown in the image viewer.
  
• If Customer is empty or incorrect, select the correct customer. The system updates defaults and re-evaluates the delivery address (see Delivery address matching).
  
• Correct Delivery Address, Order Type, Currency, and other header fields as needed.
  
• Review each line in Part Lines — verify Catalog No, quantities, units, and delivery dates.
  
• Add new lines or edit lines that were not extracted correctly.
  
• Click Next when the order details are correct.

Apply

• On Order Processing Parameters, set release and confirmation options as required.
  
• Click Finish to create the customer order.
  
• Confirm any warnings (for example, if the PO number already exists for the customer).
  
• The feature does not automatically save a customer order. You must review extracted data and click Finish to create the order.
  
• To process another document, use Restart to begin a new Scan Customer Order session. Each upload is processed independently within its own assistant session.

Example Actions

Based on extracted results, you may:

• Create a customer order from a customer PO without re-keying header and line data
  
• Select a customer manually when automatic matching is not confident enough
  
• Adjust suggested catalog numbers and quantities before order creation
  
• Add missing line items when the document layout prevented full line extraction
  
• Attach the scanned document to the created order when Document Management is installed
  
• Providing clear, legible purchase orders in supported formats such as PDF or high-resolution images helps improve extraction accuracy.

Important to Know

• Extracted data is a suggestion and must be reviewed before use
  
• The AI does not guarantee correctness or completeness
  
• The AI does not replace manual review and professional judgment
  
• Output quality depends on document quality — unclear or damaged documents may produce incomplete results
  
• Customer, delivery address, and part matches depend on your company's master data setup and must be verified manually (see How Matching Works)
  
• If no data is extracted, verify that the document is in a supported format and is legible
  
• If you see a message that the document cannot be processed, try a higher-quality scan or a different file format
  
• Missing line items on PDF documents do not always stop the process — you can complete lines manually on review
  
• Currency on the order typically comes from customer defaults; currency read from the document may not always be applied