One of South Africa’s top logistics providers has built a strong reputation on scale and operational efficiency. With a vast national footprint and over 100,000 orders processed monthly, their systems are built for speed, accuracy, and automation.
More than 90% of those orders come in through structured, EDI-integrated channels. But the final 10% — that messy, unstructured, manually processed slice — was creating real drag on performance.
When 10% causes 90% of the admin
Many of this company’s customers operate as franchise groups. Their ordering systems differ, their formats vary, and most orders from this segment arrive via email — as PDF attachments. Some of these PDFs are generated from point-of-sale systems. Others are handwritten lists scanned in-store.
Even within a single customer group, more than 10 different document layouts were in circulation.
That meant their teams had to read, interpret, and retype order details into the ERP — every day. Fields like customer name, order number, and VAT total had to be manually extracted from documents that weren’t built for automation.
At the volumes they operate, this was more than just tedious — it was costly, risky, and hard to scale.
The company partnered with Moditar to find a better way.
Enter Vision AI: Automation that adapts
Instead of adding more staff or forcing customers to change how they send orders, the solution was to automate the interpretation of those documents using Moditar’s Vision AI engine.
With Vision AI, each incoming PDF is treated like a conversation. The AI is told what to look for — no matter where or how it appears on the page:
- “Find the customer name.”
- “Get the order number.”
- “Extract the total including VAT.”
The AI processes documents in text or image format and returns clean, structured data. Because there are no templates or pre-configured layouts, the system works across formats — even if the look and feel of the order changes.
To ensure accuracy, each document is assigned a certainty score. Anything under a predefined confidence level (for example, 99%) is flagged for manual review. Everything else is safely posted to the ERP.
What changed?
With Vision AI in place, the manual processing burden dropped dramatically.
- Orders are processed faster and with fewer errors.
- Staff can focus on exception handling, not retyping.
- Customers continue using their existing formats without disruption.
- The business gains consistent data without adding operational strain.
All of this is achieved using Moditar’s Utopia platform as the engine behind the scenes — which means it fits neatly into existing integration workflows and ERP structures.
Why this matters
This use case isn’t unique. Almost every mid-to-enterprise business receives documents that don’t fit cleanly into EDI or structured API flows — whether it’s purchase orders, invoices, delivery notes, or even claims and credit requests.
Manually processing those documents is inefficient. Forcing customers to conform to rigid templates rarely works. Vision AI gives companies a smarter third option: automation that adapts to messy, real-world inputs.
And when that automation ties directly into your ERP, you gain speed, visibility, and control.
Still dealing with unstructured PDFs, scanned orders, or emailed documents? There’s a better way. Let’s talk about how we can help.