Op-ed written by Claudia Rees, Creative HQ Senior Innovation Specialist.
What a week inside a sprint actually looks like (and why it matters for the public sector)
I’ve spent enough time in the public sector to know what it feels like when a project drags. When one person ends up owning a piece of work that really needs a team. When you’re trying to get people aligned across different parts of an organisation, and the best you can manage is a series of meetings where everyone’s half-present. When the people inside the organisation aren’t trusted to solve it, and instead the answer comes from above or from an external consultant.
After years working in New Zealand’s public health sector and the UK’s NHS, joining Creative HQ at the start of February has refuelled my belief in an organisation’s ability to change. Coming from health, which is complex, resource-constrained, and full of problems that don’t have clean edges, I wasn’t sure how any of that experience would translate when I first stepped into this new role. Especially when I couldn’t quite explain to former colleagues what I actually did now.
What is a ‘sprint’? How is that relevant or useful for the public sector?
Then I experienced one.
The challenge
The organisation I worked with runs a major capital programme, the kind that has to deliver more over the coming decade while getting more out of every dollar. It needed to find real efficiencies without letting service performance slip. Those efficiencies had to add up to a real number against a real baseline, and they had to survive scrutiny from a board, a regulator, and the people who’d be asked to deliver them.
That’s the backdrop to a programme we’re running to find those efficiencies and build the capability to keep finding them. A target like that can’t simply be cascaded down, and no single team owns the answer. They have to come from across the business: from how work is planned, how it’s specified, how it’s procured and delivered, and from doing some things differently altogether.
The organisation knew it couldn’t spend months on a traditional discovery exercise for every piece of this. It needed structured, fast, evidence-based work that drew on the knowledge already inside the business. The sprint I’m writing about here was one of those pieces, an opportunity identification sprint designed to surface and size where credible efficiencies could come from, and see how they connected across the organisation rather than letting each get solved in isolation.
What happened in the room
Before the week started
Before the sprint, our team worked with the organisation to design the search. We built a framework of lenses that followed the cost lifecycle, each one a different angle on where cost is created and where it could be reduced: how decisions get made about what to build, how things get specified and designed, how work gets procured, how it gets delivered, and how assets get maintained over their life. Each lens had its own brief, its own set of questions, and its own subject matter experts.
We mapped where the money actually goes across the portfolio. We ran structured interviews with SMEs to identify the right source material: lessons learned, business cases, asset plans, feedback registers. That told us where to look for solutions and what to ask.
By the time the sprint week began, we had a clearly defined problem, a structured search, and an evidence base the team could work from.
Building a shared picture
The sprint team was deliberately built. People from across the business, chosen for the mix of skills and personalities they brought, rather than just their technical knowledge. Some were natural creators, full of ideas. Some were hackers, sharper at pulling those ideas apart and asking the harder questions. Some were heroes, the connectors who could land an idea with the rest of the organisation. Some were hustlers, the people who’d move the conversation forward and support the team to actually get things done after the week finished.
A room full of one personality type produces a thin output. The week needed people who’d disagree with each other in useful ways. It also needed a mix of seniority, teams and time in the organisation. That’s how you surface the opportunities that don’t normally get airtime, rather than only the ones the loudest voices already raise.
Bringing the right mix of people into the room
The team then brought in subject matter experts, not to sit in the room permanently, but through structured interviews at the points where their input was most needed. There are focused ways to get the insights you need without the conversation being dominated by the loudest technical voice, and without requiring every expert to be present all week.
Using AI as part of the method
This was the piece that helped us crunch through more work in a week than you’d think possible. Alongside the human team, we used AI agents to read thousands of pages of internal documents at a scale no sprint team could manage manually.
The agents didn’t decide anything. They surfaced insights and helped the room tie thoughts together. The sprint team’s job was to interrogate what came back: where the agent had picked up something helpful, where it had misread context, where it had missed something important, and where two seemingly separate findings were actually the same opportunity.
Instead of beginning with a blank wall and asking the team to generate from scratch, the team began with a structured set of candidates and spent its energy on the harder work: sizing, sequencing, and stress-testing.
Diagnoses before solutions
The most useful discipline during the week was resisting the urge to jump to solutions. Things that looked like one problem turned out to be two or three problems with different mechanisms, different owners, and different fixes. A finding about how work gets approved, for instance, can sit on top of two completely separate issues: one about the system that captures the data, and one about who owns the decision. Treat them as the same problem and the fix won’t land.
The sprint kept pulling things apart before putting them back together. It’s slower in the moment and faster in the end.
Bringing in the right expertise
Subject matter experts weren’t sitting in the room permanently. They were brought in through focused interviews and review sessions at the points where their input was most useful. There are structured ways to get the insights you need without the conversation being dominated by the loudest technical voice, and without requiring every expert to be present all week.
Designing the portfolio
By the back half of the week, the team moved from validating opportunities to shaping a portfolio. They sized each opportunity in ranges, mapped it to where the money actually sat, identified what kind of efficiency it was, and worked out what should happen next. Some opportunities were ready to be acted on. Some needed more investigation. Some needed a sprint of their own. Some needed to be parked.
Keeping leaders in the loop
Throughout the week, we were in frequent contact with senior leaders. They weren’t doing the solving. They were being kept across progress, building confidence in their team’s work, and validating the outputs at key moments. They set direction and then trusted their people to work through the detail.
What came out of it
By Friday, the team had produced a prioritised shortlist of opportunities, each one with a credible size, an evidence base, an indication of where the saving sat in the business, and a recommended next step. They had a portfolio that could survive scrutiny and a clear story to tell about how the organisation could meet its target.
In most organisations, that work would take months. Often it ends up as a consultant report that nobody owns. Here, the people who’d built it were the people who’d be champions in the business for taking it forward.
What I took away
Having a week forced decisions that might have been circling for months.
External facilitation meant the team could focus on the substance rather than managing the dynamics.
The structure was carefully designed. Every session had a purpose. The programming moved between exploration and decision-making at the right moments. There was always a point mid-week where it felt like nothing would come together, and then it did.
Using AI as part of the method changed what was possible. The agents didn’t replace the team but instead expanded its reach. The discipline of treating agent outputs as inputs to be tested, not answers to be accepted, kept the human judgement where it belonged.
The mix of people in the room mattered as much as their expertise. The discipline of separating diagnoses from solutions mattered more than the speed of generating ideas.
Alongside the tangible outputs, the team built relationships, understanding, and shared language that didn’t exist before the week started. They developed skills in structured problem-solving. They left with confidence in each other and in the work they’d built together.
We didn’t solve anything. Their team did. We provided the method, the facilitation, and the conditions. They brought the knowledge, the commitment, and the solutions.
Coming from health, I can think of dozens of scenarios where this approach would make a real difference. I’m sure it extends across the public sector more widely too: working out where efficiencies can credibly come from, redesigning a service, aligning teams that have drifted apart. The expertise is almost always already in the room. The evidence is almost always already in the documents. What’s usually missing is the structure and the space to find it.
