Deloitte maps six ERP trends as CIOs rethink core systems

Enterprise resource planning systems were supposed to be the stable layer. You implemented once, endured the consultants, then ran payroll and procurement for a decade. That deal is breaking. Deloitte has published a trend map for 2026 and beyond that treats ERP less as a fixed asset and more as a live capability that must absorb artificial intelligence, composability and continuous change without a forklift every five years.

The shift matters because ERP touches every function. Finance, supply chain, human resources and manufacturing all depend on the same transactional spine. When that spine cannot adapt quickly, the organisation cannot adapt quickly. The six trends Deloitte names describe how CIOs are trying to make core systems elastic enough to support experimentation elsewhere.

What the trends describe

Deloitte's list does not prescribe a single architecture. It describes six forces already reshaping how enterprises buy, build and operate their central systems.

Artificial intelligence moves from bolt-on to embedded. Early ERP AI was a separate module you licensed and rarely used. The new pattern embeds machine learning into procurement recommendations, invoice matching and demand forecasting as default behaviour. The system suggests, the human approves or overrides, and the model learns from the override. That feedback loop only works if the ERP vendor has designed the interface for it and the organisation has trained people to treat suggestions as hypotheses rather than edicts.

Composability replaces monolithic suites. The traditional ERP sale was a single vendor for every module. Composability lets you keep SAP for finance, swap in Workday for HR and plug a specialist tool into procurement, all talking through APIs. The trade-off is integration risk. You need a team that understands data contracts and can debug when one module's update breaks another's workflow. The upside is speed. You can replace one component without re-implementing the entire stack.

Cloud migration continues but with more nuance. Most large enterprises have moved ERP to the cloud or are mid-migration. The nuance is which workloads stay on-premises for latency, compliance or cost reasons. Deloitte's trend acknowledges hybrid as the steady state rather than a temporary compromise. That means CIOs must manage two operational models and ensure reporting works across both.

Data governance becomes a product capability, not an afterthought. Regulations such as the European Union's Digital Operational Resilience Act and Australia's Privacy Act amendments have made data lineage and auditability non-negotiable. ERP vendors are responding by building governance into the data model itself, so every transaction carries metadata about who touched it, when and under what policy. The alternative, bolting governance onto a system designed without it, is expensive and fragile.

Continuous updates replace big-bang upgrades. The old ERP rhythm was upgrade every three to five years, test for six months, go live over a weekend and hope nothing broke. Cloud ERP vendors now push updates quarterly or monthly. That cadence forces a new discipline. You cannot test every workflow before each release, so you need automated regression testing and a culture that treats small production issues as normal rather than catastrophic.

Skills shift from configuration to orchestration. ERP administrators used to spend their time in configuration screens, setting up approval hierarchies and chart-of-accounts mappings. The new job is orchestrating integrations, monitoring API health and training business users to self-serve through low-code tools. That requires a different hiring profile. You need people who understand both the business process and the technical contract between systems.

What this means for L&D

Learning and development teams that treat ERP as someone else's problem will find themselves behind. Every trend Deloitte names has a capability implication.

If AI is embedded, your people need to understand when to trust a recommendation and when to escalate. That is not a technical skill. It is judgement about data quality, model limitations and the cost of a false positive versus a false negative. Training cannot be a one-hour e-learning module. It needs to be scenario-based, with real examples of where the model has failed in your organisation and what the human did to catch it.

If your architecture is composable, your IT team needs to think in terms of contracts and dependencies, not just features. That is a shift from "Does this module do what we need?" to "Does this module expose the data and events we need to connect it to everything else?" L&D can support that shift by running workshops on API design, data modelling and integration patterns, using your own ERP stack as the case study.

If updates are continuous, your training cannot be a big event every three years. You need a system for notifying users of changes, explaining why the change happened and giving them a way to practise the new workflow before it goes live. That system is part of your ERP operating model, not a separate L&D initiative.

The governance trend is the hardest to train because it is about accountability, not mechanics. People need to understand that every data entry is a compliance artefact. That understanding does not come from a policy document. It comes from showing someone a real audit trail, walking through what went wrong when the trail was incomplete and making it clear that governance is not overhead but the thing that keeps the organisation out of court.

The skills shift from configuration to orchestration is an L&D opportunity. You can build a learning path that takes an ERP administrator from "I know how to set up an approval workflow" to "I know how to design an integration that survives an update". That path includes API fundamentals, event-driven architecture, monitoring and incident response. It also includes softer capabilities such as translating a business requirement into a technical specification and explaining to a vendor why their proposed integration will not work in your environment.

Deloitte's six trends are not predictions. They are descriptions of what is already happening in organisations that have decided their ERP cannot be static. The question for L&D is whether your training model can keep up with a system that no longer stands still.

Sources:

• CIO