Shortly after 7 a.m. in the shared inbox: one order as a PDF attachment, one as free text in the email body ("same as last week, but 20 instead of 12"), a photo of a handwritten note and a call-off from a blanket order. Someone has to read each of them, understand them and enter them in the ERP as a sales order. To automate order entry means: software reads these orders, recognizes customer, items, quantities and dates, checks them against the master data and books the order into the ERP. Inside sales only deals with the exceptions.

Key takeaways

• Order entry is the transfer of all order data into the ERP as a sales order. In technical distribution, this is still often done by hand from emails and PDFs.

• Automatic order entry with AI reads orders in any format, translates customer descriptions and old article numbers, and adds surcharges such as freight or packaging.

• EDI and web shops remain the best solution for structured orders. AI closes the gap for everything that arrives as free-form email.

• The ERP remains the leading system: prices and calculation stay there, and every order is checked before booking.

• With Mercura, the first workflows are usually live in under four weeks.

What is order entry?

Order entry is the step in which a customer order is created in the ERP or merchandise management system with all relevant data: customer, delivery address, items, quantities, units, prices, delivery date and surcharges. Only then can warehouse, shipping and accounting work with the order.

It is the first part of order processing. Order confirmation, picking, delivery and invoice follow. Errors at entry carry through every following step: an incorrectly entered item is picked incorrectly, delivered incorrectly and invoiced incorrectly. Clean entry, conversely, means fewer complaints and higher customer satisfaction.

There are basically three ways an order gets into the system:

• Manual: An employee reads the order and types it into the ERP. This manual order entry is still widespread in distribution.

• Structured digital: The order arrives as a data record via EDI, web shop or customer portal.

• Automatic from documents: Software reads emails, PDFs or scans and creates the order from them. This automated order entry is the core of digital order entry.

Why is order entry so time-consuming in technical distribution?

Order entry in technical distribution is time-consuming because many orders do not arrive as a clean data record but as email, PDF, photo or free text in the customer's own language. Before booking, someone has to understand the order and translate it into your own article numbers.

Typical hurdles from daily work:

• Customer language instead of article number: "10 pcs. elbow 45 degree 110 gray" instead of an article number.

• Customer-specific or outdated numbers: The customer orders with their own material number or with an article number that has not existed since the last range cleanup.

• References to earlier transactions: "as per quote of 12.09." or "same as last delivery".

• Delivery and pickup details: Job site address, contact on site, requested date "week 46, early".

• Surcharges: Freight, packaging or minimum order surcharges have to be added as separate line items.

• Routing: The order has to be assigned to the right branch and the responsible inside sales rep.

On top of that come call-offs from framework agreements, which must be assigned to the right blanket order. A quick calculation: if you enter orders manually, need ten minutes per order and enter a hundred orders a day, you tie up a good 16 working hours a day, about two full-time employees, just for retyping.

How does automatic order entry with AI work?

Automatic order entry with AI works like this: the software picks up orders from the shared inbox, recognizes customer and line items, matches customer descriptions to your own items, adds surcharges, flags uncertainties and books the order into the ERP after review. Inside sales only handles the flagged lines.

This is how it runs at Mercura, an AI software for quote and order automation in technical sales:

The most important difference from classic text recognition: AI-based entry understands the order in terms of content. Behind it is artificial intelligence, more precisely machine learning and language models. It does not just read characters at fixed positions on a form, but recognizes that "110 elbow 45° gray" belongs to a specific item in your range, even if the customer writes it slightly differently every time.

Example: An email order becomes a sales order

Example (anonymized): This is what a typical order from an installation company to a plumbing and HVAC distributor looks like:

Subject: Order Schulstraße project

Hello all,

please deliver to job site Schulstraße 14, contact Mr. K. Delivery Thursday morning.

• 12 x HT pipe DN 50 2 m

• 20 elbows 87° DN 110 gray, same as last time

• 4 pcs. 4711-220 (your old number)

• 1 can lubricant

Thanks and regards

This results in the following draft order:

The inside sales rep reviews two flagged points instead of retyping six lines and looking up numbers. Then the order is booked into the ERP.

OCR, EDI or AI: Which method fits which order?

For structured orders from large customers, EDI is the best solution. For customers with always identical forms, classic OCR with templates may be enough. For all free-form orders by email, PDF or photo, AI-based entry is the right path. Most distributors need a combination.

You will find detailed comparisons at EDI or AI? and OCR or AI?

How does the order get into the ERP?

The order gets into the ERP through an interface: in real time via API, for SAP for example via OData, or as a file via SFTP, for SAP for example as an IDoc file. The ERP remains the leading system. Prices and calculation are not moved to a second piece of software.

Mercura is connected to SAP S/4HANA, SAP ECC, SAP Business One, Microsoft Dynamics 365, Salesforce, proALPHA and UniTrade, among others, and to other systems via an API or SFTP. Important for IT leads: the connection runs only via routes your IT approves and controls. The data is stored in Europe, every customer has a separate tenant, and customer data does not train external AI models. Mercura is ISO/IEC 27001 certified and works in compliance with GDPR.

We describe the SAP variant in detail at Automate SAP order entry, and all systems on the Integrations page.

Order processing software or AI order entry: What is the difference?

Software for order processing and order fulfillment, usually part of the ERP or merchandise management system, manages the order after it is in the system: confirmation, warehouse, shipping, invoice. AI order entry sits in front of it and makes sure that unstructured orders arrive in the system as a clean order in the first place.

So it replaces neither the ERP nor an order management system. If you are looking for order processing software, first clarify where the bottleneck is: in fulfillment after booking or in entry before it. If it is in entry, a new ERP is rarely the answer. Usually, all that is missing is the automatic bridge from the inbox to the existing system.

Where are the limits of automatic order entry?

The limits are where information in the order is missing or contradictory: software cannot reliably identify an item the customer does not describe clearly, and it cannot set terms that have not been agreed. Such cases land with inside sales flagged, not guessed.

Typical borderline cases:

• Unclear details: "the usual" or "as discussed" without any reference.

• New customers without master data: The customer has to be created in the ERP first.

• Very poor scans or illegible handwriting: They are flagged, not guessed.

• Phone orders: They do not land in the inbox and remain a separate channel.

• Very low volume: With only a few orders a day, rollout often does not pay off.

To be honest: if your most important customers already order via EDI and the rest consists of a few emails, expanding EDI or the web shop may be the more economical investment. AI order entry pays off when a large share of orders arrives as free-form email.

How do you roll out automatic order entry?

Roll out order entry automation step by step: first one branch or one inbox, real orders from the last few weeks as test material, clear rules for surcharges and approval, then expansion. That way you see early how many lines run through without correction before you change any processes.

A proven sequence:

1. Measure the starting point: number of incoming orders per day, time per order, share by email, PDF and fax.

2. Collect test material: real orders including difficult cases such as photos, free text and old article numbers.

3. Define rules: When is freight charged, which packaging applies, who is responsible for which branch?

4. Define the review step: Who approves flagged lines, and how fast?

5. Go live and evaluate: share of lines without correction, time per order, backlog at the end of the day.

6. Expand: more inboxes, branches or, in addition, quote processing.

How do you recognize a good order entry solution?

A good solution recognizes orders in any format without a template per customer, matches customer descriptions and old article numbers to your own range, flags uncertainty visibly and books directly into existing systems such as your ERP. Test this with your own orders, not with the vendor's demo data.

With Mercura, the technical test with your own data usually runs within two working days, and the first workflows are usually live in under four weeks, without prior training. Mercura is used by more than 80 distributors and manufacturers in Europe and the US, rolled out across 115 locations. You will find details about the product on the Order processing page. If you also want to automate quotes, start at Automate quote creation.

Test it with your own orders

Bring a real order: in 30 minutes we show you how Mercura handles it in your process. Book a meeting

Conclusion

Order entry in technical distribution is often still manual work because orders arrive as free-form email, PDF or photo. AI can read these orders, match customer descriptions and old numbers to your own range, add surcharges and book the order into your existing ERP. EDI and web shops remain sensible for structured customers. For everything else, a test with real orders decides how much retyping you can spare your inside sales team.

Further reading

• Order processing: emailed orders into your ERP automatically

• OCR or AI? Automatically Capture Orders and Requests

• Sales Order Automation FAQ: 16 Questions Answered for Distributors and Manufacturers

• Customer story: Sanitär-Heinze (plumbing and HVAC wholesale)

• Mercura vs. Workist: BoQs, GAEB and Order Entry Compared

• Mercura vs. turian: AI for BoQs, Quotes and Orders

Frequently asked questions

What is order entry?

Order entry is the transfer of a customer order with all its data into the ERP or merchandise management system as an order: customer, delivery address, items, quantities, prices, dates and surcharges. It is the first step of order processing.

How does order entry work?

Order entry is done either manually by retyping, in a structured way via EDI, web shop or customer portal, or automatically, with software reading emails, PDFs and scans and creating an order from them that is checked before booking.

What is the difference between order entry and order processing?

Order entry is the starting point: the order is created in the system. Order processing covers the entire further path with confirmation, picking, delivery and invoice.

Can AI also capture orders as PDF, scan or photo?

Yes. Mercura reads orders from email text, PDF, Excel, scans and photos. If a document is barely legible or a detail is unclear, the line is flagged and reviewed by inside sales.

Do we still need EDI?

Yes, EDI remains sensible for customers with high, structured order volumes. AI order entry complements EDI for all orders that arrive by email and cannot be handled via EDI.

What happens if the AI does not recognize an item?

The line is flagged, not guessed. The sales rep selects the right item before the order is booked into the ERP.

How long does rollout take?

With Mercura, a technical test with your own data is usually possible within two working days. The first workflows are usually live in under four weeks, without prior training.