JD Edwards users find practical value in AI through measurable workflows

For many JD Edwards users, the question is no longer whether AI agents are interesting, but whether they can be justified in hard business terms. Kevin Van Horn of ERP Suites argues that the answer depends less on the technology itself than on the process it is attached to. In his view, AI only begins to generate a return when it is introduced into a measurable, repetitive workflow with a clear owner and a baseline that shows what improvement looks like.

That shift, he says, re...

flects a broader change in the market. JD Edwards leaders have moved beyond curiosity and are now pressing vendors and consultants for evidence that AI can solve real operational problems. The most compelling use cases are not abstract experiments, but tasks where staff spend time rekeying data, chasing approvals, checking statuses or handling exceptions. Van Horn says those conditions are most common in high-volume, rules-driven work, where digital workers can take over routine activity and free people for higher-value decisions.

He draws a sharp distinction between a useful pilot and a technology exercise. A genuine pilot, he says, starts with a defined business problem rather than with a desire to “use AI”. The process should be understood, the process owner identified and the metric agreed in advance. Without those three elements, he warns, organisations are likely to end up with an expensive science project rather than a business case.

Among the strongest starting points, according to Van Horn, is accounts payable. He points to the long-running effort to automate invoice handling, arguing that AI-powered digital workers represent the next stage in that evolution. Rather than relying on optical character recognition and other older methods, he says modern systems can capture invoices from email, portals and cloud storage, validate the information and load it into JD Edwards with far less manual intervention. The metrics that matter most are not just days payable outstanding or discounts captured, but also invoice throughput, first-time success rates and the amount of human touch required.

Order entry is another area where he sees immediate potential. Even where electronic data interchange already handles much of the flow, many organisations still rely on people to process thousands of manual sales orders each month. Van Horn says AI can accelerate that work by reading documents such as PDFs or spreadsheets, extracting key details and proposing how an order should be created according to business rules. In a high-volume sales environment, he argues, speed and accuracy in order capture have a direct effect on customer responsiveness and downstream fulfilment.

Procurement also stands out as a strong candidate. Van Horn describes it as a process rich in exceptions, approvals and data points that need constant monitoring. He says digital workers can help from supplier onboarding through to purchase order tracking, delivery monitoring and even retainage release, all while pulling information together in real time. That visibility, he suggests, is particularly valuable because procurement teams often spend too much time looking for information that a well-designed automation layer could surface automatically.

He is equally enthusiastic about finance close. Although closing the books is monthly rather than continuous, he says its difficulty lies in coordination, timing and exception handling. Digital workers can monitor status across tasks, identify anomalies in trial balances and reconcilations, and focus attention on the small minority of items that need action. The goal, he says, is to remove the burden of scanning through large volumes of correct data so that finance teams can concentrate on the exceptions that actually affect the close.

When it comes to ROI, Van Horn says leaders should be careful not to overstate labour savings. He cautions against treating automation as a headcount-reduction exercise, arguing that the real value is in capacity, resilience and labour efficiency. A digital worker may reduce overtime, help a business absorb growth without increasing headcount and lessen the disruption caused by turnover, but it should be framed as a way to improve productivity and employee experience rather than as a simple replacement strategy.

That framing matters, he says, because employee satisfaction is often overlooked in ROI calculations. Removing repetitive, low-value work can make roles more appealing and help organisations keep experienced staff, particularly in areas where institutional knowledge is hard to replace. In his view, that is one of the most practical benefits of automation: it allows people to spend more time on judgement, service and analysis rather than routine administration.

Van Horn also warns against anchoring business cases to demo results or best-case assumptions. Instead, he says organisations should use average baseline data over a meaningful period and include the real costs of implementation, support and change management. Conservative assumptions are safer, he argues, because they reduce the risk of disappointment and make it easier to defend the investment internally. In his experience, the actual outcomes often exceed the original forecast once the automation is embedded properly.

The broader message is simple. For JD Edwards customers, the best AI opportunities are not the flashiest ones, but the ones that remove friction from familiar processes. According to ERP Suites, that means starting with the work that is repetitive, measurable and painful today, then building the business case from there.

Source: Noah Wire Services