- 19
- February
A universal problem — why bother setting up long approval chains if no one actually reviews them and just clicks through? By the time someone discovers the budget was charged to the wrong project, the money has already been spent. The issue is that the requester selects the wrong project budget — and it is a project within their own department. What makes it even harder — even AI cannot determine whether the selection was wrong.
A Real-World Scenario
Picture this — you are an officer managing multiple projects within the same department. Suppose you oversee Project A, Project B, and Project C. All projects are under the same department and share identical expense categories: compensation, travel, project development, operations, and so on.
One day you need to claim travel expenses for Project B. You open the system, select "travel expenses" from the dropdown, and choose the budget — but accidentally select Project A instead of Project B.
What happens next?
- Does the system check if there is sufficient budget? — Yes, because Project A also has a travel budget. It passes.
- Sent for approval through the chain of command — Every step is correct because it is your own project. The approver is your department head. It passes.
- Does the approver know it is wrong? — There is no way to tell because everything looks correct. Travel expenses are a legitimate item, the budget is sufficient, and the approval chain is correct.
- The error is only discovered when... — You notice "why has Project A's money decreased?" and then discover it was your own selection error.
What is worse — if no one ever goes back to check, nobody may ever know there was a disbursement to the wrong project. The money is already spent, the budget is already reduced, and the project budget figures are completely distorted.
Why This Problem Is So Difficult to Fix
Logically, the ERP system should catch this, right? But look at what the system has already done:
| What the System Already Does | Status | Does It Catch the Error? |
|---|---|---|
| Access restriction — user sees only their own projects | Correct ✓ | No, because the user has access to both projects |
| Approval route — sent to the right person through multiple reviewers | Correct ✓ | No, because the approver only sees "travel expenses from Project A" — nothing looks abnormal |
| Budget validation — checks sufficient funds | Correct ✓ | No, because the budget is actually sufficient |
| Real-time dashboard — project owner can view anytime | Correct ✓ | Helps "detect early" but does not "prevent" |
| All projects share identical expense categories | — | Cross-validation by expense type is impossible |
In short: everything is "correct" except the project selection — this is a human error the system cannot detect because there is no data to indicate "this travel claim should come from Project B, not A."
What ERP Can Help With (But Cannot Prevent 100%)
Although it cannot fully prevent the problem, ERP offers approaches that help reduce the chance of errors:
| Approach | Method | Limitation |
|---|---|---|
| Link activities to budget | User selects the activity/plan first, and the system automatically pulls the Budget Code | Requires detailed activity planning, which many organizations do not do |
| Clear UI design | Display the project name prominently with different colors throughout the form | Reduces mistakes but does not prevent them |
| Summary before submit | Display a message "You are claiming [expense] from [Project X]" for confirmation | Users may click through without reading |
| Monitoring dashboard | Project owner sees real-time spending totals | Helps detect early but does not prevent |
| Easy reversal | Has a process for budget transfer/reversal when errors are found | Corrects after the fact but does not prevent |
As you can see, every approach has limitations. No method prevents the problem 100% — because the issue is not a system error but "human selection error" in a situation where every option appears correct.
Can AI Help?
This is an interesting question — many people think AI could catch this error. But think carefully:
- All projects share identical expenses — there is no pattern for AI to detect. Travel expenses from Project A look exactly the same as those from Project B.
- The user has access to both projects — nothing unusual from an access rights perspective.
- No external data can tell — AI has no way of knowing that "this travel expense was for a Project B activity, not Project A" if that information was not specified from the start.
The only thing AI can help with is: analyzing spending patterns per project. If spending deviates from the average — for example, Project A has unusually high travel expenses this month — it could trigger an alert for review. But that is still "detection after the fact," not "prevention."
Once Found, It Is Easy to Fix — But Why Do Users Resist Fixing It?
Once you know the wrong project was selected, the fix in ERP is straightforward:
- Cancel the incorrect document — on the date it was discovered
- Create a new document — selecting the correct project, also on the discovery date
- Done — it does not affect previous budget figures even if that period was already closed, because it is a new entry in the current month
Extremely simple — but the problem is that users do not want to do it.
Why? Because users want the "original date" — they want the expense to appear in the "original month" when the transaction was made. This is simply not possible if that month's books are already closed. And this is not a system limitation — it is a fundamental accounting principle that prohibits retroactive changes to entries in a closed period.
What Users Often Request vs. What Can Be Done
| Want the original date | Not possible — cancellation and re-creation happen on the current date |
| Want it in the original month | Not possible — if that month's books are already closed |
| Want the system to fix it automatically | Not possible — the system does not know which project is "correct" |
| Want the support team to fix it | Cannot help — because only the user knows which project is correct |
Ultimately, the only person who can fix this is the user themselves — because they are the only one who knows which project the expense should belong to. The support team, system administrators, and even executives cannot make that decision.
The correct and simplest approach is: cancel on the day you find the error, re-create on the same day — done. It does not affect closed month budgets, no need to reopen periods, no need to retroactively adjust figures.
"The Old System Could Fix It, Why Can't ERP?"
This question comes up frequently — users often argue that the old system (or another system) allowed them to open approved documents and simply change the project code. Why can't ERP do the same?
The answer is: ERP can do it, but intentionally does not allow it.
Because directly editing an approved document destroys three critical things:
| What Is Destroyed | Impact |
|---|---|
| Audit Trail | No evidence that the document was modified, who changed it, when, or what was changed — if auditors ask, there are no answers |
| Approval chain | The document was approved based on one set of data, but the actual data has changed — the approver never saw the modified information |
| Credibility of figures | If anyone can edit approved documents, the report figures lose credibility — because there is no way to know how many times they were changed |
A system that "allows editing" sounds convenient, but it is really a system without Internal Control — which is precisely what the organization invested in ERP to fix in the first place.
The "cancel and re-create" process is not much harder than "open and edit," but the benefits are enormous — complete evidence of why the original document was cancelled, what the new document references, and that it went through proper re-approval.
And importantly — if the system "magically fixes" things every time, users never learn. They pick the wrong project, ask for a fix, pick wrong again next time, and ask for another fix — an endless loop. The "cancel and re-create" process may seem slightly more cumbersome, but that small inconvenience is precisely what makes people remember to be careful, because if they choose wrong again, they will have to go through the process once more.
An Important Lesson — Not Every Problem Can Be Solved by a System
This is a critical lesson for every organization about to purchase or implement an ERP system:
- ERP is not a magic cure-all — it helps tremendously in many areas, but some problems simply cannot be solved by any system.
- Some problems are human errors that the system can only "reduce the likelihood" and "help detect sooner" — not prevent 100%.
- Fixing it is easier than you think — just cancel and re-create. Not complicated, but you must accept that the date will differ from the original.
- What matters more than the system is good work processes + human diligence + regular reviews.
- A good ERP system accepts that people make mistakes and is designed to make "correction easy" rather than "prevent everything."
A good ERP system is not one that prevents everything, but one that when mistakes happen, they can be detected quickly and corrected easily.
— Saeree ERP Team
Conclusion — How Much Can the System Help?
- 100% prevention is impossible — if the user has access to multiple projects with identical expense categories, the system has no way of knowing the wrong one was selected.
- Can reduce the chance of errors — through clear UI design, pre-submission summaries, and linking activities to budgets.
- Can detect problems faster — through real-time dashboards and project-level spending reviews.
- Can correct conveniently — through budget transfer/reversal processes with full audit trails.
- The rest is up to people — diligence, regular verification, and an organizational culture that pays attention to detail.
If your organization is facing challenges with budget control, disbursement approval, or needs a system that helps reduce errors and detect problems faster, you can schedule a Demo or contact the consulting team to discuss the work processes that best fit your organization.
