Picking a winner is harder than it looks

Five shortlisted problem solving initiatives, and there can only be one winner. If Sam and I are honest, we had no clear idea who would take it.

I’d been to the Tilley Awards before, in a previous role at South Yorkshire Police, so I knew what to expect - experienced practitioners sharing great work on problems affecting victims, witnesses and communities. What I didn’t expect was how hard it would be to separate this year’s finalists.

The range of projects was amazing. The Police Service of Northern Ireland introduced thresholding panels to triage lower-risk cases, cutting the number of registered sex offenders without a risk management plan by 42%. South Yorkshire Police tackled serious violence in Sheffield’s night-time economy, with no stabbings there since November 2024 and policing costs down by £2.16m. The Met used the Serious Crime Act to charge firearms brokers even when no weapon is ever recovered. And Humberside Police’s work on Grimsby’s West Marsh estate cut violent crime by 35% and anti-social behaviour by 61%.

Any of them could have won, but the prize went to Greater Manchester Police for Operation Wage. Rather than chasing the volume of fraud, they treated it as a problem of repeat victims, focusing on the vulnerable people being targeted again and again. Fraud fell by 55–75% across key districts, and around £2.8m was recovered. Congratulations to DI Stacey Shannon, DCI Charlotte Whalley and the team.

We were sponsors at the three-day conference at St George’s Park, and across two days of workshops, around 80 people joined us to work through a complex policing problem. Both sessions were full, and about half stayed behind afterwards for a demo of our wider policing work. We didn’t start with an AI solution and work backwards, we introduced it step by step to see where it would fit and add value.

We ended up discovering that AI is useful for retrieving information and surfacing insight, putting the right material in front of an officer. How to respond is still the officer’s call, because that depends on investigative judgement and local knowledge.

Professor Nick Tilley joined us on day one, and at the start of day two spoke positively about how the workshop showed AI supporting the problem-solving process. That was good to hear, and the conversations with practitioners were just as valuable.

One theme kept coming up over the three days - how much effort it takes to search through historic problem-solving plans. Those plans record the problem, the intervention and what happened next, and they’re a goldmine if you have a week spare to read them.

Practitioners talked about how summarisation could surface relevant past work (what worked, what didn’t, and in what context) without someone trawling records by hand. That could give officers time back, and something genuinely useful to draw on when they face a new problem.

The workshop and our stand led to conversations with people interested in an early proof of concept, those are at an early stage, and we’re looking at what might be possible with interested forces and partners. But the real value for me was hearing directly from the people doing this work, and testing our thinking against their experience.

We came home with plenty to follow up, and far fewer rubber ducks and cookbooks than we left with. And as a Nottingham Forest fan, I’ll admit there was something very satisfying about spending three days at St George’s Park watching footballers actually put the ball in the back of the net. It’s been a while.

‍