By Phil Abbott
ERP systems integrate across your entire business, handling everything from quoting and sales orders to inventory management and financial reporting. Every user interacts with a different slice of it. The sales team knows their workflows. Finance knows theirs. Stock management operates in its own world. Each department does its bit brilliantly, then passes documents along the chain – but cross that boundary and suddenly you’re in unfamiliar territory.
This creates a silo problem. It’s remarkably silent, but pervasive across organisations using ERP. The trouble emerges when you need to step outside your silo. When a manager or CFO wants a broader view, they suddenly need to access different parts of the system and gain insights from modules they’ve never touched.
This is where AI is making a significant impact. It’s not replacing ERP systems, it’s unlocking what’s always been there, making the full power of these platforms accessible to everyone who needs it.
The insight problem
Traditionally, getting shared insights between departments has been challenging. You had to know where to look, which reports to run and how to interpret the data. Forecasting sales meant spending hours looking for trends in historical data. I spoke with a customer recently who had been trying to forecast, and business had been either through the roof or through the floor, with nothing in between. Nobody had any idea what was going on. These are exactly the situations where AI can genuinely assist, though I’m careful to say ‘assist’ rather than ‘replace’.
AI-enabled systems give you insights rather than requiring you to find them. AI’s ability to surface insights, identify trends and flag potential issues is becoming remarkably sophisticated. AI algorithms can now run much faster than before and are being applied to ERP in genuinely useful ways, improving usability while extracting more meaningful information from data.
The human element remains critical
Here’s what I always emphasise to clients: AI will assist with forecasting, trend identification and insight generation, but it still needs a human to look over the results and do something about it. That human oversight isn’t a weakness of AI – it’s an essential part of making it work properly.
An emerging skillset involves the person who can take AI-generated insights from the ERP, analyse them critically, verify their accuracy and then implement business decisions based on that analysis. But there’s another critical side to this skill: gathering the correct information in the first place.
I regularly have customers who want specific reports: “I want to know when this happened and what this did.” My response is always: “Great, we can build that. But where are you recording when this happened?” Because AI can only work with the data it’s given. If you’re servicing devices and you want to know turnaround times, you need to timestamp when the device arrived, when work started, when work completed.
This is the foundational truth: if date fields aren’t populated, if warranty periods aren’t recorded, then the ERP can’t automatically track warranty expiration dates. And if the data isn’t there, AI can’t extract insights from it.
The training challenge
When someone leaves a role, they typically train their replacement, but they can’t train them properly because they only know their specific part of the system. Take supplier returns: someone in quality receives a faulty device and needs to send it back. Simple enough, except there’s actually a complex decision tree: Will we get a replacement unit? Will they give us a credit note? Will they take it back without credit? Each scenario requires different handling in the ERP.
These decisions have huge implications. If you forget to invoice a customer for a service contract you’re delivering, you lose that revenue while still doing all the work. This is precisely where AI notifications can prevent revenue leakage and human errors.
I genuinely believe AI is going to build more intuitive signposting into ERP systems. We’ll see features that ask, “are you sure you wanted to do that?” when something seems unusual or incorrect, guiding users towards the correct processes. We’re moving from systems that passively accept whatever users input, to intelligent platforms that actively prevent errors and surface opportunities.
AI in ERP isn’t about replacing human judgment. It’s about making these vast, complex systems work the way they were always supposed to; accessible to everyone who needs them, not just the experts who’ve spent years learning the intricacies. When insights surface automatically, when silos become transparent and when every user can make informed decisions regardless of which module they typically work in, that’s when ERP systems truly deliver on their promise.
Phil Abbott is Director at ABS Limited. For more information visit: abslimited.co.uk
Read other recent UK Manufacturing news: https://uk-manufacturing-online.co.uk/category/news/
