3 Minute read

Building an ERP with AI

Building an ERP with AI? Cedar Bay explores why working software is only the start, and what it takes to run everyday business processes reliably.

Dan Boulton

2nd Nov 2026

Can AI Build Your ERP? The Last Mile Problem

By Dan Boulton | Cedar Bay

AI is changing the way businesses develop software. Tools such as Claude and ChatGPT are making it possible to build applications and automate processes at a fraction of the time and cost previously required.

For businesses frustrated with their existing systems, the prospect of building exactly what they want is understandably attractive.

But just because AI makes it easier to develop software, does that mean building your own ERP is a sensible business decision?

Building an ERP with AI: Why it gets complex

In logistics and distribution, the last mile is often the most difficult and expensive part of the journey.

You can move goods efficiently across continents, through ports and distribution centres, but getting that final parcel to the customer’s doorstep can introduce unexpected complexity.

I think developing an ERP using AI presents a similar challenge.

AI might get you remarkably close, remarkably quickly. You could have functioning screens, workflows and reports in a fraction of the traditional development time.

But that final stretch is where the real complexity sits.

Take something as simple as processing a sales order. Beyond capturing the order, the system needs to manage stock availability, credit limits, VAT, allocations, invoicing and financial postings.

What happens when an order is partially shipped, cancelled or returned? How does that affect stock, customer balances and financial reporting?

These aren’t unusual exceptions. They’re everyday business processes.

Building an application that performs a task is very different from building a platform that reliably runs your business.

And the danger is that you may not discover what you’ve underestimated until you’ve already invested considerable time and money.

Build your own ERP or extend an established platform?

Established ERP platforms have spent years developing the transactional controls and business logic that organisations rely on.

Rather than using AI to recreate these foundations, why not use it to develop functionality that makes your business different?

An established ERP can provide the core processes, while AI creates opportunities to automate tasks, improve reporting and develop specialist applications.

For me, that’s a much more sensible approach.

AI may make building software easier, but it doesn’t necessarily make owning and operating that software any less complex.

In my next article, I’ll explore what I believe is an even bigger consideration: the security, compliance and financial risks of developing your own ERP.

Let us help you transform your business

Get a Free Demo