AI in ERP: Can we Start Looking to the Future?
For years, ERP systems have told businesses what happened yesterday: recording orders, tracking stock, processing invoices and producing reports. Useful, certainly, but fundamentally backward-looking.
Does Artificial Intelligence have the potential to change that?
The key shift is not that ERP has suddenly become intelligent, though some may claim otherwise. It is becoming better at identifying patterns, highlighting risks and anticipating what may happen next: evolving from a system of record into a system of decision support. Gartner calls this ERX (Enterprise Resource Execution).
For SMEs, strong opportunities lie in forecasting and planning. Demand and inventory forecasts still rely heavily on human judgement, spreadsheets and historical averages—often, at best, an educated guess. AI can analyse much larger datasets, spot overlooked trends and update forecasts continuously as circumstances change.
The practical benefits include lower excess inventory, better service levels, earlier visibility of emerging demand and greater confidence in cashflow projections. Managers can spend less time gathering information and more time deciding what to do with it.
To Proactive from Reactive?
Can AI shift management from being reactive to proactive? Conceivably, rather than reporting a problem after it occurs, systems can warn that one is likely—a production bottleneck, stock shortage or unexpected change in customer demand—before it affects the business.
However, SMEs should be realistic. The biggest challenge is not necessarily the technology itself, but the quality of the underlying data. Inaccurate stock records, duplicate customer data, inconsistent processes and disconnected systems all undermine AI-generated insights. Put simply, AI cannot create reliable answers from unreliable information.
It’s all about the data…
“AI-ready” should therefore mean “data-ready”. Before investing heavily in AI-enabled ERP, businesses should prioritise process discipline, data quality and system integration. These topics may not be exciting, but they often determine whether the outcome is confidence and success, or fear and disappointment.
Any ERP upgrade, replacement or simply refresh should absolutely have a stream dedicated to preparing at the data. That may involve cleansing and organising existing data, and should definitely consider how data will be created going forward to ensure it is clean and useable from its inception.
Simple Guidelines:
- Avoid free text fields. It is tempting, for speed, to allow users to enter their own data, but it is a recipe for confusion. For example, is it “LTD.”, “Limited”, “LIMITED”, or “Ltd.”?
- Structure them instead. Ensure you have policy and means of providing a structured alternative. For instance: a postcode (or Companies House) lookup for addresses and names, and segmented item descriptions using selection boxes for defined values.
- Minimise mandatory fields. Make sure they are only used for data that is genuinely needed operationally, for compliance, or to drive system behaviour.
- Create clear data ownership. “If everybody owns it, nobody owns it”. Every data area should have an owner to be responsible for its quality. For instance, Customers – Sales Manager, Suppliers – Procurement Manager, Products – Engineering/Product Management, Chart of Accounts – Finance, Employees – HR
- Limit those who create data. Not all users have the time or desire to ensure data is clean. Limit those who create it and ensure those who have to are guided by the system. A simple approval process to ensure sign off by the data owners is valuable.
- Standardise Units of Measure. Make sure EA is used consistently, is it cm or mm?
- Consult regularly with the data users. Are your team able to report adequately? Can they find what they want easily, and it is fit for purpose?
- Have regular data quality reviews. and track challenges such as duplicate customers and suppliers, missing fields, inactive items, etc.
The Power of Information Flow Analysis
AI will almost certainly become a standard ERP feature rather than a specialist add-on. Managers are likely to use natural language to question business systems and receive recommendations, instead of searching through reports and dashboards.
The greatest advantage will not necessarily go to businesses with the most AI, but to those with the best data, clearest processes and strongest understanding of the decisions they want to improve.
That is where the real value lies, though even then, Black Swan events remain possible: no person or system foresaw Covid, the Suez Canal blockage, or the wars in Ukraine and Iran, all of which profoundly affected supply chains and costs.
Strong data. Better decisions. Greater confidence.
AI can only be as effective as the information it relies on. Establishing clear data ownership, consistent master data standards and robust ERP processes lays the foundation for reliable forecasting, planning and operational control. To discuss your ERP strategy, data governance approach or digital transformation programme, contact Gradient Transforming Limited and discover how we can help your business create lasting value from its systems and data
