Probabilistic tech improves police response times in North Wales 2026

Key Points

• North Wales Police and Quantum Dice have reported on the findings from the PROTECT proof of concept which used probabilistic computing for planning the deployment of emergency response vehicles.
• In a constrained, simulated environment, average response times were reduced by 40%, down to just 4.82 minutes compared to a baseline station approach, while the proportion of the highest priority (P0) calls that could be responded to within 10 minutes increased from 58.8% to 75.6%.
• The geographic ward coverage in the model also increased from 39.3% to 64.3%.
• In another 24-hour simulation of 13 emergencies, the minimum average response time achieved was 7.76 minutes, exceeding the project target of around 10 minutes although, according to organizers, more validation is required before putting it into operation.
• The research received support from the UK Policing National Science and Innovation Board via their Test and Learn Program with a prototype tool combining optimization, simulation and visualization provided.
• Related media coverage states that the underlying methodology was that of quantum annealing with an overall objective of the force being the reduction of response time goals in half.

Wales (Wales Times) September 2026 — The two organisations have disclosed findings from Project PROTECT, a proof-of-technology initiative that applied probabilistic computing to the forward deployment of emergency-response vehicles, reporting substantial reductions in modelled response times and improved coverage of high-priority demand.

What are the Project PROTECT results for response times and coverage?

As reported by UK Authority’s coverage of the announcement, the initiative demonstrated that probabilistic optimisation can materially improve forward-deployment planning while adhering to operational constraints considered critical in policing. In a constrained modelled deployment scenario—where one vehicle was retained at each of five station wards and the placement of ten surplus vehicles was optimised—average response times were reduced to 4.82 minutes, representing an approximate 40% improvement over a station-based baseline.

The same modelling showed the proportion of highest-priority (P0) demand reachable within 10 minutes increasing from 58.8% to 75.6%, with geographic ward coverage rising from 39.3% to 64.3%. Beyond that experiment, Quantum Dice built a dynamic optimisation and simulation environment that recalculated vehicle placements as incidents occurred and vehicles became available; in a 24-hour simulation covering 13 urgent incidents, all were responded to under work-rest scheduling constraints, with the best-performing configuration achieving an average response time of 7.76 minutes.

How does probabilistic computing support police vehicle deployment?

According to The Quantum Insider’s press-release summary, Project PROTECT evaluated how Quantum Dice’s probabilistic processor can support emergency-response vehicle deployment planning, with the project delivering a working decision-support prototype that combines optimisation, simulation and visualisation capabilities. George Dunlop, Co-founder and Director of Partnerships at Quantum Dice, said:

“Project PROTECT demonstrates that advanced optimisation technologies can deliver meaningful operational insights for public-sector organisations today. By leveraging our UK-made probabilistic processor, we have shown how sovereign UK computing capability can support critical public-safety missions and provide a practical pathway to scalable deployment.”

Alistair Hughes, Advanced Analytics and AI lead at North Wales Police, said:

“Optimising Forward deployment remains one of most challenging problems in policing. Project PROTECT allowed us to evaluate a probabilistic computing approach addressing optimisation targets and also critical operational constraints that are common across policing environments. The results provided valuable evidence, demonstrating the potential benefits of further development and future deployment.”

The force operates across a geographically diverse region where response planning must balance speed, rural coverage, station resilience, vehicle availability, officer welfare and operational practicality, the announcement noted.

What do police and innovation officials say about the pilot?

A spokesperson for the Office of the Police and Crime Commissioner for North Wales (OPCCSA) said:

“Projects such as PROTECT help us move beyond discussions about emerging technologies and assess their potential in the context of real operational requirements. Through the NSIB Test and Learn Programme, this work has generated evidence on the application of advanced optimisation techniques to a nationally relevant policing challenge, while highlighting opportunities that may also be relevant across wider public services. It demonstrates the value of collaboration between policing and innovative technology partners in exploring future capabilities.”

The proof-of-technology was supported through the Test and Learn Programme run by the UK Policing National Science and Innovation Board (NSIB), which has positioned the work as generating evidence on advanced optimisation for a nationally relevant policing challenge. Organisers emphasised that results are based on historical demand data and that further validation would be required before any operational deployment.

How does this relate to wider quantum and optimisation efforts in policing?

Reporting by Quantum Zeitgeist describes the broader context in which North Wales Police has been pursuing quantum annealing to address the “ten-minute challenge”, aiming to halve response-time targets while maintaining coverage for around half a million residents across rural and urban areas. That coverage quotes Alistair Hughes explaining that measuring time from call receipt rather than dispatch effectively tightened targets, and that hybrid quantum-classical systems were used to handle data preparation and optimisation in an integrated workflow.

The same reporting notes that a standalone application now informs vehicle placement decisions that previously required extensive manual analysis, and that the force is exploring future integration of machine learning to anticipate demand fluctuations. While Project PROTECT focuses on probabilistic computing, the overlapping narrative in sector coverage underscores a wider push to apply advanced optimisation—whether probabilistic or quantum-annealing based—to resource deployment across emergency services.

Background to the development

Project PROTECT follows a series of UK policing initiatives exploring artificial intelligence and advanced analytics to reduce administrative burdens and improve response outcomes. In August 2026, the Home Office announced a national AI call-routing system for 101 non-emergency calls intended to slash waiting times and free officer time by redirecting misdirected calls to other agencies such as councils or the NHS. Forces including South Yorkshire Police and Cumbria Police have begun using AI triage on 101 calls, with government estimates suggesting potential annual savings of around £8.5 million for the sector.

Within that landscape, Project PROTECT represents a parallel strand focused on the physical deployment of response vehicles rather than call handling. Supported by the NSIB Test and Learn Programme, the project produced a prototype decision-support tool and modelled outcomes that indicate significant potential improvements in response-time metrics and coverage, while explicitly flagging the need for further validation before live operational use.

Prediction: how this could affect police forces and the public

If subsequent validation confirms the modelled gains, probabilistic optimisation could affect police forces by enabling more granular, data-driven forward deployment that balances speed of response with geographic equity and officer welfare constraints. For the public, particularly in rural and semi-rural areas, the most tangible effect would be a higher likelihood of high-priority incidents being reached within target windows, alongside more consistent coverage across wards.

For senior leaders and operational planners, the technology offers a decision-support layer that can be rerun as daily availability changes, potentially reducing reliance on static rosters and manual heat-map analysis. However, the organisers’ caution that results are based on historical data and require further testing implies that any rollout would likely proceed through additional pilots, integration work with existing command-and-control systems, and governance reviews to ensure transparency and accountability in how deployment recommendations are generated and used.