PARTNER

Why gridlocked agencies should start with understanding, not new tech

As AI moves from experimentation to real-world adoption, public sector leaders are navigating shifts that go beyond technology alone, such as the foundations their systems are built on, to how decisions are made, to how risk and trust are understood.

For public agencies that feel gridlocked, the starting point is a clearer understanding of the existing systems. Image: Canva

This article is part of a Thoughtworks thought leadership series on unlocking public sector AI innovation. You can click the author's byline to read other articles by him.

For public agencies that feel gridlocked, the starting point is rarely a new platform or tool.

It begins with a clearer understanding of the existing systems, including applications and the value streams governing how changes are requested, built, approved, and released.

These are socio-technical systems, shaped as much by process and policy as by code.

Without this understanding, modernisation risks becoming expensive technology substitution, rather than meaningful transformation.

The hidden cost of centralisation

Many large institutions are still navigating the consequences of earlier architectural decisions.

Highly centralised systems make change both slower and more expensive, as every modification

*James Lewis is a programmer and Director at Thoughtworks based in the UK. Image: Thoughtworks*

requires coordination, approval and access to shared components. Teams begin to queue for access, and delivery slows.

It's not due to a lack of capability, but the structure.

This trade-off is particularly relevant for public agencies. Centralisation brings control and cost benefits but often comes at the expense of speed.

Decentralised approaches can restore autonomy and pace, but this only works when teams maintain strong engineering discipline and governance.

A pattern across the region

Across conversations with Chief Information Officers (CIOs) and chief architects at government agencies in Singapore and Australia, one constraint continues to surface regardless of agency, scale or mandate.

Legacy systems are still doing most of the work, and often hold everything else back.

Agencies are finding that without a solid foundation, artificial intelligence (AI) and other emerging technologies are difficult to apply safely and at scale.

There is no single starting point. While some organisations have embarked on multi-year modernisation efforts, others are still trying to define what modernisation should mean for them.

The one clear pattern that emerges is that the organisations best positioned to capitalise on what comes next are not the ones reacting to AI today, but the ones who have already invested in how their software is built, deployed and operated.

What 'modern' actually looks like

These organisations are not defined by a single technology, but by a set of consistent traits.

They use microservices architectures where appropriate, and prioritise platform engineering and developer experience.

They are increasingly comfortable operating in cloud-native environments, with stronger visibility into how systems behave in production.

Over time, these institutions begin to resemble digital-native organisations, even if they were never designed that way.

In practice, this might involve funding shorter-lived initiatives, revisiting approval thresholds, or creating space for teams to test ideas before committing to full-scale programmes.

This disciplined engineering approach is a critical enabler when AI or other advanced technologies are introduced.

Prioritising what really matters

Once that understanding is in place, prioritisation becomes grounded.

Some systems are costly to maintain or nearing the end of their viable life. Others consume disproportionate resources for low-value outcomes.

Most importantly, government agencies must focus on systems that directly shape citizen experience, such as services that reduce friction, improve access, and make interactions simpler and more effective.

Focusing on these areas first creates momentum and makes the case for continued investment clearer, both internally and externally.

Modernisation as a long game

Modernisation is not a short-term effort. Most organisations should expect it to span several years, requiring sustained executive sponsorship and careful sequencing of work.

While AI can accelerate parts of this journey, they do not replace the need for foundational capabilities.

Foundations such as continuous delivery, automated testing, reliable build pipelines and alignment between systems and business domains make systems easier to change, safer to operate and resilient to future disruption.

There is no single path through this transformation. Different organisations will sequence their efforts differently, based on their starting point and constraints.

One principle remains consistent: to take advantage of the next generation of technology, agencies need systems that can evolve safely and continuously.

Put simply, they need to be tall enough to get on the ride. With these foundational capabilities in place, agencies are better positioned to act decisively as new technologies arrive.

------------------------

James Lewis is a Thoughtworks Distinguished Engineer and Director based in the UK, internationally recognised for software architecture and its intersection with organisational design and lean product development. He co‑authored the original 2014 definition of microservices with Martin Fowler, helping catalyse the industry’s move to independently deployable services. He serves on the team behind the Thoughtworks Technology Radar and regularly advises executive teams on aligning technology strategy and organisational structures to accelerate delivery.