Automated Order & Schedule Processing

About the Client

ACS Industries is a customer-focused industrial organization that manages purchase orders, delivery forecasts, and schedules across a growing customer base. With more than 30 customers involved in the process, the client required a scalable and automated approach to handle inbound order information and integrate it efficiently with its IFS ERP system.

The Challenge

Diverse formats — PDFs, Excel files, scanned images, tables pasted into the email body, and .eml
attachments.
No shared template — every customer’s document layout is different across 30+ customers.
Image documents — scanned and picture-based documents must be read by OCR before any data can be
used.
Regional formats — US vs. European dates, and European decimals (e.g. "335,000" means 335, not
335,000).
Non-order content — totals rows, footers, addresses and header fragments can look like real order lines.
IFS is strict — a blank date or a wrong-typed field causes IFS to reject the entire record (deserialization
errors).
Per-customer quirks — highlight-colour rules, missing labels, multi-section spreadsheets, page-break splits,
and cumulative-quantity reconstruction.
Routing — correctly separating orders from schedules and saving each to the right destination for posting.

The Objective

The objective was to automate the processing of customer orders and delivery schedules by eliminating manual data entry into IFS. The solution needed to handle documents across different formats and customer layouts, accurately extract and transform business data, and reliably post structured records into IFS at scale—reducing errors and freeing teams for higher-value activities.

The Solution

Central extraction engine — an Azure Function with a dedicated handler per customer, mapping each
document to the IFS order or schedule format.
OCR — Azure Document Intelligence converts scanned and image documents into machine-readable text
and tables.
Intake & routing — Power Automate captures each email and routes the document by type.
Delivery — Azure Blob Storage and Azure Data Factory pipelines post the finished records into IFS.
Deterministic validation — valid-part-number gates, totals/phantom-line filtering, unit and date
normalization, date-based line typing, and IFS-safe handling of typed fields.
Iterative customer resolutions — each customer validated against real documents and client QA, resolving
image-OCR, regional numbers, missing labels, phantom lines and cumulative reconstruction one by one.