You cannot reshape a public service if you cannot change the systems that run it.
That should be obvious. Yet transformation programmes can still treat the legacy estate as a technical problem to address somewhere beneath the ambition. The conversation starts with new digital experiences, better use of data and AI-enabled services. The systems that must support those ambitions become somebody else’s workstream.
That separation is a mistake. The legacy estate helps determine what a service can deliver, how quickly it can adapt and how much each improvement costs.
The problem is what you cannot change, confidently
An old system is not automatically a bad system. The harder problem is a service whose behaviour nobody can fully explain, with dependencies that are poorly understood and changes that are difficult to test.
Replacing it does not make those uncertainties disappear. They re-emerge in migration, integration and the discovery of operational rules that were never documented.
In public services, those rules matter. They can determine whether an application is processed correctly, a payment reaches its recipient or staff can access the information they need.
Engineering-led modernisation starts by understanding that reality. Leaders need evidence about what can change, what depends on it and how the service will continue to operate.
Deferring change also needs scrutiny. A service can remain operational while the skills to maintain it disappear, dependencies become harder to support and each improvement takes longer. Keeping it running is not sufficient evidence that its future is secure.
Diagnose. Prioritise. Modernise.
This is the discipline needed to modernise.
Diagnose means examining the working service – its code, data, integrations, infrastructure and the manual processes keeping it moving. It also means listening to the people who operate and use it.
Prioritise means deciding which changes will improve outcomes, reduce exposure and unlock further progress. The most visible technology problem is not necessarily the most valuable one to solve first.
Modernise means engineering a practical route forward, with testing, security and operational readiness built into delivery.
This gives leaders a basis for committing to change and teams a route they can execute.
Time to value belongs in the engineering decisions
Public services should not have to wait until a transformation programme finishes before seeing improvements.
Time to value depends on how work is sequenced. Removing a manual handoff, resolving a data problem or making deployments repeatable can deliver benefits while creating better conditions for the next change.
Our work on the UK Ship Register illustrates why the foundations matter. Alongside the digital service, we addressed legacy data migration, validation and integration with GOV.UK services. Registration processing fell from more than 60 minutes to 8–17 minutes. The improvement reached the operation behind the screen.
The lesson is to connect engineering decisions to service outcomes from the outset, then organise delivery so value can emerge along the way.
Safe to change is a lasting capability
Keeping a critical service running today is essential. Making it easier to improve tomorrow is part of the same responsibility.
That requires more than a new platform. It requires systems that are understood, reliable delivery practices and knowledge transferred to the teams responsible for the service. Every programme should leave public sector teams better equipped to make the next change.
Public sector transformation ambitions will remain constrained if changing the underlying estate continues to be treated as too difficult to confront.
Make those systems safe to change and more ambitious public services become possible.
