The operating model is the constraint.

When a platform is introduced into an operating model that has not changed, the enterprise does not simplify. It gains another interface onto the same unresolved way of working.

Complexity increases because the old path is rarely retired. Legacy process survives alongside the new system, and the organisation now maintains both — with integration, reconciliation and exception handling between them.

The business case assumed replacement. The operating model delivered addition.

An operating model is not an organisation chart. It is the working answer to five questions: who decides, who owns, how work moves, what is measured, and who is accountable for the result.

Those five answers determine whether a capability creates value. A platform can make a process faster; only the operating model determines whether the faster process is the right one, whether anyone owns its performance, and whether a decision follows from what it reveals.

Where those answers are ambiguous, technology inherits the ambiguity and industrialises it.

Enterprise architecture is often positioned as a technology discipline concerned with systems, standards and estates. Used well, it is the connection between business strategy and the way the enterprise actually runs.

That connection is the useful part. It shows where a declared strategic priority has no process owner, where a critical decision has no defined path, and where two business units are being asked to deliver one outcome without a shared measure.

Architecture that describes only the technology estate cannot see any of that — and cannot explain why investment is not converting into performance.

Major technology decisions are commitments about how the enterprise will operate for years. Made before operating-model clarity, they lock in ambiguity at the point when it is most expensive to revisit.

Clarity first does not mean redesigning the organisation before anything can start. It means naming process owners, decision rights and outcome measures for the area the technology is intended to serve.

That work is unglamorous. It also determines how fast decisions travel, where they stop, and who answers for the outcome.

Operating-model constraints announce themselves in recognisable ways. Decisions escalate because no one below the executive has the authority to make them. A process improves in one function and degrades in another. A programme reports green while the business outcome does not move.

Each of these is read as a delivery or governance failure. Each is usually a symptom of unowned work.

The diagnostic value is high: constraints of this kind are visible within weeks, and they are correctable without a new platform.