Vertical AI will expose the limits of generic ERP

Generic ERP was built to standardise the basics: finance, procurement, inventory and reporting. That model still works. But AI is exposing where it stops working.

An AI assistant can summarise a purchase order or flag an overdue invoice. It becomes even far more valuable when it understands why a machine stoppage is relevant, how a regulated batch must be handled, or which service-level commitment is at risk.

That is why vertical AI will expose the limits of generic ERP.

Context is the missing ingredient

AI needs more than clean data. It needs context: the terminology, workflows, controls and trade-offs that define how a particular industry operates.

A generic ERP AI tool may recognise that a stock level has fallen below a threshold. A manufacturing-specific tool can connect that signal to production schedules, maintenance history, supplier lead times and quality constraints.

The difference is not cosmetic.

It determines whether AI produces an interesting alert or a decision that someone can act on.

Our analysis of Acumatica’s practical AI push explored the growing emphasis on embedding AI in day-to-day ERP tasks. That is the right direction. But its usefulness depends on whether the system understands the operating environment around those tasks.

Generic data creates generic answers

ERP vendors have traditionally designed platforms to serve multiple sectors with configurable modules. That flexibility has advantages, particularly for mid-market organisations that need broad functionality without the cost of a bespoke platform.

Yet configuration alone has limits.

An AI system trained on generic finance, supply chain and customer data may struggle when confronted with industry-specific terms, rules and exceptions. In healthcare, it must understand governance and patient safety. In construction, it needs to account for contracts, site conditions and subcontractor dependencies. In manufacturing, it must recognise the links between quality, production, maintenance and traceability.

Without this context, AI can make confident recommendations that are operationally irrelevant.

Or worse, wrong.

Industry clouds may hold the advantage

Industry clouds offer a possible answer because they combine sector-specific data models, workflows and controls with cloud-based ERP capabilities. As we noted in our coverage of Infor’s industry-cloud strategy, the strategic value lies in making enterprise platforms more relevant to local and vertical operational realities.

But vertical AI is not simply a vendor feature to switch on.

Businesses still need good master data, clear ownership of processes, and firm controls over permissions, exceptions, and audit trails. AI cannot repair weak processes. It can only scale their weaknesses faster.

The next ERP battleground will not be who has the most AI agents.

It will be who has the deepest understanding of the work those agents are being asked to do.