AI assistant “Cecil” could save police 23,000 hours a year, trial finds

Humberside Police and Coeus Software have published the results of an artificial intelligence (AI) trial that could save more than 23,000 hours of officer time annually, according to new modelling.

The “Test and Learn” initiative, known as Project Moriarty, explored how AI could help frontline officers access accurate and timely procedural guidance while out in the field. Funded through the National Police Chiefs’ Council (NPCC) and National Science and Innovation Board (NSIB), the trial supports the Home Office Police Efficiency and Collaboration Programme and its efforts to keep “officers in communities, equipped with the information and tools they need to do their job, today and in the future.”

The project combined Coeus Software’s secure PoliceBox mobile platform with a new AI-powered assistant named Cecil, providing officers with verified procedural advice drawn from official police policies, guidance and legislation, but with no reliance on public or external data sources.

Addressing Experience Gaps on the Frontline

According to Humberside Police, the initiative responds to one of policing’s most pressing challenges: the erosion of experience among frontline officers. As of March 2024, 35 percent of police officers in England and Wales had fewer than five years’ service, compared with just 14 percent in 2016, while voluntary resignations also reached record levels.

To help bridge that gap, Coeus developed Cecil, described as a digital mentor within the PoliceBox app. Officers can ask conversational questions such as “I’ve arrived at a road traffic accident” and receive relevant procedural guidance in seconds. The system uses a Retrieval-Augmented Generation (RAG) AI engine to analyse authorised materials and deliver answers.

During the trial, officers rated Cecil highly for accuracy, relevance and ease of use:

• Accuracy: 4.6 / 5 (92 percent rated 4 or higher)
• Relevance: 4.5 / 5 (88 percent rated 4 or higher)
• Ease of use: 4.6 / 5 (84 percent rated 4 or higher)
• Usefulness: 4.4 / 5 (86 percent rated 4 or higher)

Response times averaged between five and twenty seconds, with half of all queries answered within seven seconds.

If adopted force-wide, Humberside Police estimates that Cecil could save more than 23,000 officer hours per year – equivalent to returning several full-time officers to the frontline.

“The Project Moriarty trial with Humberside Police showed how the Cecil AI assistant can deliver real-time guidance to officers in the field. When combined with the power of PoliceBox to digitise frontline policing, it has the potential to transform efficiency and enhance community engagement,” said Simon Hall, CEO of Coeus Software.

“The original Chief Constable Moriarty helped define policing professionalism and procedural discipline. Our ‘Cecil’ continues that legacy for the digital age.”

AI as a Tool for Confidence and Consistency

Colin Dring, Innovation Manager at Humberside Police, said Project Moriarty has shown how AI can enhance officer confidence and consistency in real time.

“By placing guidance directly in officers’ hands, we’re helping them make better decisions for the communities they serve,” he said.

One detective constable who took part in the trial added: “This would be very useful from an investigation view as it has all the data readily accessible for officers. This would improve investigation standards if they had the information on their devices.”

Named after Chief Constable Cecil Charles Hudson Moriarty OBE, whose 1920s handbooks shaped professional policing standards, today’s digital Cecil aims to guide officers through complex operational decisions “with clarity, accuracy, and confidence.”

Based in Middlesbrough, Coeus Software develops AI-powered mobile solutions for public safety, health and local government. Its PoliceBox platform, hosted in the AWS UK Police Assured Landing Zone, enables officers to complete secure digital workflows on mobile devices while remaining “visible in their communities.”