Epicor order entry: how to stop retyping customer POs into Kinetic or Prophet 21

The job nobody planned for

Walk past the order desk at most manufacturers or distributors running Epicor and you'll see the same thing. One screen shows a customer's PO, usually a PDF from an email. The other shows Sales Order Entry in Kinetic or Order Entry in Prophet 21. Someone is reading one and typing into the other: customer, ship-to, PO number, part numbers, quantities, prices, dates.

Epicor itself calls manual order entry labor-intensive and time-consuming, and lists the forms orders still arrive in: mail, fax, email, paper, PDF and spreadsheets. That matches what we see. A few big customers send EDI. Everyone else sends whatever their own system prints.

This guide covers the ways to stop the retyping, what Epicor offers for each of its two main ERPs, and how to choose. It builds on our guide to whether to replace Epicor or automate around it. If you've decided to keep Epicor, this is the next step.

What retyping actually costs

The best public benchmark we found is older, and it comes with a caveat. In 2016, the research group APQC published Cutting the Costs of Sales Order Processing, sponsored by Esker, a company that sells order automation. It used APQC's benchmarking data from 608 organizations to compare the total cost of managing sales orders, per order placed.

Total cost to manage sales orders, per order

APQC benchmarking data, 608 organizations, 2016. Covers the whole "manage sales orders" process, not typing alone.

1. Top performers$5.11
2. Median$24.21
3. Bottom performers$40.87

Source: APQC, Cutting the Costs of Sales Order Processing, 2016 (sponsored by Esker).

Bottom performers spent almost eight times what top performers did. APQC names labor as the biggest cost driver, and it points out the hidden part: the time spent when part of an order is lost, when data entry errors are made, or when a change to an order arrives.

The same report split costs by how orders came in. Orders through paper-heavy "traditional" channels cost a median of $12.60 each. Orders through digital channels cost $4.00.

You may have seen a similar number on Epicor's site. Its ECM automation page says manual sales order entry costs between $24 and $41 per order. Epicor doesn't cite a source, but those figures line up with APQC's median and bottom-performer numbers above, which cover the whole order process, not just keying. Use them as a rough range, not as the cost of typing one PO.

Errors matter as much as time. Research on data accuracy in clinical trials offers a useful comparison. A review of 93 studies found that when people read records and copied the data out, the pooled error rate was 6.57%. Typing from a form designed for data entry had a pooled error rate of 0.29%. That's medical data, not purchase orders, but the lesson carries over: reading a document someone else designed and transcribing it is far more error-prone than plain typing. A customer PO is exactly that kind of document.

A customer PO wasn't designed for your ERP. That's why typing it in goes wrong.

Five ways orders can get into Epicor

1. EDI

Electronic Data Interchange is the oldest answer. The customer's system sends the PO as structured data, and it lands in your ERP with no typing. Epicor's EDI integrations page lists both Kinetic and Prophet 21, supports POs, invoices and advance ship notices in the common X12 and EDIFACT formats, checks documents against each trading partner's specs, and offers managed services.

Good for: large customers who already send EDI and require it. Watch out for: setting up each trading partner takes work, and most of your smaller customers won't ever send EDI.

2. An online ordering portal

Prophet 21's sales page says orders entered at the counter, from a mobile device, or online by the customer go straight into P21, and points to Epicor Commerce for online ordering.

Good for: repeat buyers who order from a catalog and are willing to log in. Watch out for: many business customers won't change how they buy. Their purchasing system sends a PDF, and that's that.

3. Epicor's document capture (Kinetic)

For orders that arrive as documents, Epicor sells Sales Order Automation, a module in its Enterprise Content Management (ECM) suite. Epicor describes three steps:

• Capture. Intelligent Data Capture pulls out the customer, order details and line items.
• Validate. It looks up data in Kinetic to confirm the customer, quote and other checks. Exceptions go to a person to resolve.
• Process. If everything checks out, the order is created in Kinetic.

Epicor's ECM automation page says its capture reads key fields without templates and supports PO capture from PDF and TIFF files, manual review and exception routing. It lists integrations with Kinetic, Vision and Microsoft Dynamics GP. Prophet 21 isn't on that list.

Good for: Kinetic shops that want to stay inside Epicor's own tools. Watch out for: the pages give no accuracy figures, so test it on your own messiest POs before you buy.

4. A partner tool (common for Prophet 21)

For Prophet 21, Epicor points to a partner. Its partner directory calls Conexiom Epicor's official partner for Eclipse and P21 order automation. The listing says Conexiom turns emailed PDFs, spreadsheets, images, text and faxes into ERP-ready data and checks it against Epicor records. In July, Conexiom announced it was expanding the partnership to cover RFQs, sales orders, AP invoices and vendor acknowledgments for distributors. Accuracy and volume figures on those pages are Conexiom's own claims.

Good for: distributors on P21 or Eclipse who want a packaged product. Watch out for: the listing names only Prophet 21 and Eclipse, so Kinetic shops should ask.

5. A custom build on Epicor's API

The last option is software built for your order desk that reads each PO, checks it against your Epicor data, and creates the order through Epicor's API. For Kinetic, Epicor's Open REST API page says anything available in the user interface is available through REST, and that security follows your existing ERP user and data permissions. It also says the API lets software work with ERP services directly "rather than screen automation."

That last point matters. Some automation tools work by copying keystrokes into ERP screens, the way a person would. When a screen changes in an upgrade or a move to the cloud, those scripts tend to break. Building on the API avoids that.

Prophet 21 is less clear from public pages. Epicor's P21 technology page describes an SDK for customizing P21 without changing its source code, but doesn't describe an order API. Ask your Epicor account team what integration options your version supports before you plan a build.

Good for: desks with many customer layouts, revised POs, and checks that are specific to your business. Watch out for: it's a project, so it needs a clear owner and a defined scope.

What users say about the do-it-yourself route

Some Epicor shops build their own workarounds. On EpiUsers, a forum for Epicor users, one Kinetic user described turning a big customer's spreadsheet into an import file for Epicor's Data Management Tool, for about 200 orders every two weeks. The user wrote that they still have to look over the orders because the customer doesn't stick to set conventions. That's one user's experience, not an Epicor statement, but it's a common story: a clever import handles one customer's format, then the format drifts.

It's also worth knowing that Epicor positions its Data Management Tool mainly for moving data from older systems into Kinetic, not for daily order intake.

What a good setup does, whichever you pick

The tool matters less than the steps. A setup that actually takes the typing off people's desks does five things:

1. Reads every format you get. PDFs, scans, emails with the order in the body, spreadsheets. If it only handles your top three customers, the desk still types the rest.
2. Checks against Epicor. Is this a real customer and ship-to? Do the part numbers match, or does the customer use their own? Does the price match the quote or price list? Is the date possible?
3. Shows a person what it found. The PO and the draft order side by side, with anything unusual flagged. The person approves or fixes it in seconds.
4. Posts through the API. Not by typing into screens.
5. Handles changes. Revised POs and change orders are where manual desks lose the most time. The setup should spot a revision and show what changed.

That's how we built order entry for a food-equipment manufacturer that supplies national grocery chains: the software reads each customer order and fills it in, and their team approves it, in the system they already run.

When not to automate

To be fair, not every desk needs this. If you get 20 orders a week from five customers, a good template and a careful person may be the cheapest answer. If one customer sends most of your volume, ask them about EDI or a portal first. And if your Epicor setup itself is the problem, with messy part numbers or price lists nobody trusts, fix that first. Automation checks orders against your ERP data, so it's only as good as that data.

The case for automating grows with three things: volume, the number of different layouts, and how much a mistake costs you in returns, credits and angry customers.

What to do this week

1. Count the orders. For one week, count the POs your desk keys in by hand, how many customers they come from, and how many layouts.
2. Sort your customers. Which few could send EDI? Which would use a portal? Which will always send PDFs?
3. Time a few orders. Pick five typical POs and five messy ones. Time them start to finish, including fixing mistakes.
4. Ask Epicor what you already own. Do you have ECM, EDI or REST access licensed? What does your P21 version support?
5. Test any tool on your worst POs. Not the clean demo file. The scanned, rotated, three-page one.
6. Read the related guides: how to automate order entry and what handles PDFs that vary.