Coeus Software worked with The Scale Factory and Amazon Web Services (AWS) to deploy a secure, police-assured cloud environment on AWS, enabling the delivery of AI-enabled frontline workflows within the UK Police Assured Landing Zone (PALZ). The engagement supported the deployment of the PoliceBox platform within PALZ, the rapid establishment of a compliant AWS environment, and the successful delivery of Project Moriarty, an AI capability designed to provide procedural guidance to frontline officers using controlled, auditable AI techniques.
Background
Coeus Software develops PoliceBox, a frontline application and platform designed to support police officers when information surfaces for action or when officers are collecting information in the field. PoliceBox provides a secure frontline entry point to multiple back-office and national systems, supports offline working, and is designed around an interoperability-first, open-API architecture rather than bespoke integration. This allows forces to configure and evolve frontline workflows without repeatedly rebuilding underlying systems.
As part of its roadmap, Coeus began developing Project Moriarty, an AI-enabled capability intended to support frontline officers with procedural guidance, particularly in complex or unfamiliar situations. The objective was to support officer judgement and consistency while maintaining strict controls around data sources, assurance and auditability.
The Challenge: Setting Up PALZ and Delivering AI at Pace
UK policing operates in a highly assured technology environment. Any platform handling police data must meet stringent security and governance requirements, including compliance with the Police Assured Landing Zone.
Coeus made a strategic decision to deploy PoliceBox on AWS to support:
- Scalable, on-demand computing
- Secure access to managed AI services such as AWS Bedrock
- Alignment with emerging national cloud patterns across policing
PALZ defines required outcomes and controls, but it does not provide a fully prescriptive implementation. Translating those requirements into a working, compliant AWS environment requires deep expertise across networking, identity, permissions and service boundaries.
At the same time, Project Moriarty introduced additional complexity. The AI capability was designed using Retrieval Augmented Generation (RAG), ensuring responses were constrained to approved national and force-specific sources rather than general internet content. This required careful configuration of AWS Bedrock within a tightly controlled environment.
While Coeus had strong product and policing-domain expertise, the organisation did not have sufficient in-house AWS experience to stand up a PALZ-compliant environment quickly enough to support live policing trials.
Partnering with The Scale Factory and AWS
Recognising the need to move quickly without compromising assurance, Coeus engaged The Scale Factory as an AWS specialist delivery partner.
The Scale Factory worked closely with Coeus to:
- Set up a fully compliant Police Assured Landing Zone on AWS
- Deploy the PoliceBox platform within that environment
- Configure secure access to AWS Bedrock for AI services
- Implement infrastructure-as-code to automate and repeat the build
AWS actively supported the engagement, recognising the strategic importance of enabling assured AI experimentation within policing. AWS funding supported The Scale Factory’s involvement, allowing the platform foundation and AI capability to be delivered in parallel.
What Coeus Software said...
We needed to establish PoliceBox in a Police Assured Landing Zone quickly, and we simply didn’t have the AWS experience in-house to do that at pace.
The Scale Factory brought exactly the specialist capability we needed, and AWS support helped us accelerate delivery without cutting corners.
The Solution: Police-Assured AI, Delivered Securely
The resulting solution provides Coeus with a fully police-assured AWS foundation, designed to be repeatable and scalable.
Project Moriarty uses Retrieval Augmented Generation to provide officers with procedural guidance based exclusively on approved sources, such as national guidance and force documentation. General internet data is deliberately excluded to reduce the risk of incorrect or inappropriate responses.
The Scale Factory played a critical role in:
- Designing a secure network architecture within PALZ
- Configuring identity and access controls for AI services
- Ensuring AWS Bedrock operated entirely within the assured boundary
- Automating the deployment so environments can be recreated consistently
By scripting the entire build, Coeus can now spin up a new, compliant backend environment in minutes, supporting future pilots and deployments without repeating foundational work.
The Results
The platform supported an NPCC NSIB-funded Test & Learn trial with Humberside Police, integrating AI-driven procedural guidance into frontline workflows.
Measured outcomes from the trial included:
- Approximately 23,000 officer hours per year saved, based on observed use cases
- Rapid delivery, with a PALZ-compliant AWS environment established in days rather than months
- Assured compliance, enabling Coeus to demonstrate a police-assured platform when engaging with forces
- Reduced AI risk through constrained, auditable RAG-based AI rather than open-ended models
The work also resolved a common barrier in the policing market: the need for a demonstrably compliant environment before forces are willing to engage seriously with new technology.
Looking ahead
With a secure, repeatable AWS foundation in place, Coeus Software is now focused on expanding the capabilities of Project Moriarty, including features such as shift summarisation, automated prompts, and deeper integration with frontline workflows.
The collaboration with The Scale Factory and AWS has enabled Coeus to move faster while maintaining the assurance standards required for policing, supporting a platform designed to scale as policing needs evolve.
“This partnership allowed us to focus on delivering value to officers, rather than building and maintaining cloud infrastructure ourselves,” said Hall. “It’s a model we expect to continue as the platform grows.”
