Our takeaways from SDinGov 2025

Rethinking systems and complexity

Day one’s keynote talk set that tone, with a scene-setting talk that highlighted how modern day government emerged during the age of enlightenment in the 18th century to deal with clear, linear problems like public health and law and order. While this model worked for transactional challenges, it struggles in the face of complexity arising from the repeated black swan events faced globally, such as the dot com bubble, financial crisis of 2008, and the Covid-19 pandemic.

James Plunkett (Kinship Works) argued that this “mechanical” approach has worked incredibly well for analytical problems, but some of today’s most pressing challenges, like housing, social care, and mental health, need a relational, people-centric approach. That means balancing left-brain efficiency (for transactional services like getting a passport) with right-brain empathy and connection (for complex, human issues).

The Ministry of Justice workshop “The chaos whisperers: reimagining the iceberg model to map systems and address complex challenges”, built on this. Using the iceberg model to show how surface events are shaped by deeper patterns, structures and mental models. Complex challenges can’t be treated as if they were merely complicated - you can’t knock them down and rebuild from scratch. Instead, the approach was to start where you are, map the system, and work iteratively. Tools like COM-B for behaviour change and a probe–sense–respond cycle were suggested as ways to navigate this complexity. The advice: “rock the boat, but stay in it” - investing time to understand systems while acting as intrapreneurs inside government to help new approaches stick.

Designing with care

Building on this recurring theme was the need to put care at the centre of service design.

Rachael Dietkus (Social Workers who Design) made a powerful case for treating care as part of the fundamental infrastructure of public services - something that should be measured as a KPI alongside efficiency. She warned that neutrality protects the status quo: if we don’t deliberately design for care, we risk creating systems that unintentionally cause harm. In other words, trauma can be designed in through confusing forms, dehumanising processes, and interactions that prioritise compliance and control over people’s needs.

Her call to action: slow down where necessary, bring empathy into every stage of the process, and work towards a new, empathy-driven model. She suggested a global code of care for digital services as a guide for how we build with dignity, not just efficiency.

The Essex County Council team also demonstrated how care shows up in practice. Their work with young people with additional needs focused on making education and training options navigable. Through ethnographic research - meeting families at home, using visual tools like the blob tree, and sharing insights through audio - they created a more human way to capture experiences and influence decision-making. They also highlighted the ethical duty to route out-of-scope insights to the right people.

Measuring what matters

Data and measurement were another thread running through the event.

Defining what ‘public design’ is can be difficult enough, let alone measuring its impact. Step forward the Public Design Evidence Review, published last July.

Kara Kane (Cabinet Office) and several of the review’s authors shared highlights from the UK Government’s comprehensive set of insights, evidence, case studies and emerging ideas on the role, value and required conditions of good design in public products and services.

Catherine Hope and Jim Montgomery (DWP) shared how performance analysts are embedded in design teams to create meaningful measures of success. Their approach, framing every piece of work with “We believe that… so if… we will see…”, ensures services are grounded in clear outcomes.

The example they shared on child maintenance services showed how thoughtful metrics not only encouraged digital uptake but also safeguarded against excluding people without digital access. It was a reminder that success measures shape services - and need to reflect real-world diversity. Adding these to every Jira story helps the team to focus on the desired impact of the feature.

AI, responsibility and inclusion

AI was inevitably part of the conversation, but the message was clear: it’s not a silver bullet.

Steph Wright (Scottish AI Alliance) warned that introducing AI into flawed systems only amplifies their problems. The people most affected when things go wrong are those with the least power and agency.

She pushed for a shift in focus: not just on technical accuracy, but on fairness, justice and responsibility. Who is accountable when AI fails? How do we design systems that include marginalised groups from the outset? And crucially, should we always be using AI at all?

Trust in AI isn’t a given, and nor should it be. Examples shared, from biased fraud detection algorithms to opaque immigration decision-making and the Post Office scandal, show how damaging misplaced trust in technology can be. The takeaway: move slow and fix things. Design for failure as well as success, and set a people-first north star. When you design for the margins, you design for everyone and build trust in the process.

Final reflections

Across the sessions, there was a strong sense that public services are at a crossroads. The old efficiency-driven model is breaking down, trust is eroding, and complexity is everywhere.

But there’s also hope. By designing for care, mapping systems honestly, measuring what really matters, and treating AI with responsibility, service designers can help shape a more human future for public services.

A huge thank you to the organisers and speakers for creating such an open and thought-provoking event. We’ll be carrying these insights forward in our work, and continuing to explore how design can support trust, dignity, and care in public services.