Skip to content
All insights

General

Your ERP Wasn't Designed for Your Workflow — Customize the Software, or Retrain Your People?

By Shanshan "Sasha" Ma

Off-the-shelf ERP is designed for the average company, and you are not the average company. Customize the software until it fits, or retrain your people until they fit it — both charge a price nobody quotes upfront. Here is what each one really costs, and the third option.

"Our ERP does the standard process fine. Our process isn't standard."

That sentence explains most ERP disappointment. Off-the-shelf ERP is designed for the average company. It is the sum of a hundred implementations, distilled into a process that fits the most buyers reasonably well. You are not the average company — and the parts of your operation that make you money are usually the parts that deviate from the average.

When the software and the work disagree, there are only two answers on the table: customize the software until it fits, or change the work until it fits the software. Both have a real cost, and neither one is the one quoted to you at signing.

The customization trap

Customizing an ERP to match your workflow is the answer most operations leaders reach for, because it is the only answer that respects how the business actually runs. It also creates four problems that compound.

  • You are now maintaining software. Every customization is code your team owns — through every upgrade, patch, and vendor release. The system becomes a branch you have to keep merging by hand.
  • The upgrade tax. Vendors ship updates for the standard product. The further you have drifted, the more expensive each upgrade becomes, until the honest decision is to skip upgrades entirely — and slowly fall off support.
  • The knowledge walks out the door. Customizations are usually understood by the consultant who built them and the two people who were there at the time. Nobody else can safely touch them.
  • The process freezes. The customization encodes how you worked on the day it was built. Change the workflow and the software becomes the obstacle again — except now it is expensive to change.

None of this is a reason to avoid configuration. It is a reason to stop using the ERP as the place your workflow lives.

What adoption actually costs

The other answer is to make your team work the way the software expects. That bill arrives in three parts.

Training and retraining. People learn the system, then learn it again at each release. The cost is not the training day; it is the weeks of slower work on either side of it.

Workarounds. Whatever the system doesn't handle gets handled beside it — a spreadsheet, a shared inbox, a personal list. The ERP is technically the system of record, and the real process lives in the gap around it. Now you have two versions of the truth and a person who reconciles them.

Loss of the thing that made you good. The reason a customer buys from you instead of a competitor is often the judgment your team applies at the exceptions. Adoption projects quietly remove that room to maneuver, because the software can't represent it. You standardize your way into being average.

And the adoption cost is recurring. Every new hire learns the workaround, not the process. Every process change reopens the argument with the software.

Why the fit problem doesn't go away

ERP is excellent at what it was built for: authoritative records, financial control, the transaction of the standard case.

The problem is that the work rarely arrives as a clean transaction. It arrives as an email, an attachment, a revised date, a phone call, a document someone scanned, and a question about which version is current. Somebody has to gather that, reconcile it against the record, notice what's missing, and decide what happens next — and then go make it happen.

That work has no home in the ERP, because it isn't a transaction. It is the context around a transaction. So it lands on a person, and the person becomes the integration layer between the systems you bought to remove exactly that job.

The off-the-shelf fit problem is not a software defect. It is a category gap.

Where LevelOneAI wins

We don't replace your ERP, and we don't ask it to become something it isn't. We take on the part it was never designed to hold: the state of a piece of work while it is still moving.

  • No customization project. Your ERP stays standard, upgradable, and supported. We sit beside it rather than inside it.
  • No forced adoption. Your team doesn't have to be retrained into a new process. The queue meets them where the work already arrives — in email, in documents, in the systems they know.
  • The context gets assembled for you. The order, the messages, the attachment, the last time this happened, the current status — gathered per work item instead of reconstructed by hand.
  • Exceptions surface before they cost you. What's missing, what's late, what conflicts, what the next step is, and who owns it — visible in one queue rather than in somebody's memory.
  • Your judgment stays yours. Consequential decisions keep a human approval, and approved changes are recorded back where the record belongs. We automate the gathering, not the deciding.

And because the work item is the unit rather than the transaction, the same approach covers the processes that span systems — order to cash, supplier exceptions, service intake — where the ERP only ever sees its own half.

The economics are the point. Instead of a multi-year customization program that freezes your process and taxes every upgrade, you keep the software that works and add the layer that carries the work between it and your people.

The test

Pick the workflow that gives your team the most trouble. Ask two questions: how much of it does the ERP actually do, and how much of it is a person carrying information between systems?

If the second answer is larger than the first, the problem isn't your ERP. It's that nobody built the layer where that work should live.