Temporary accommodation has become a major fiscal and administrative burden for English local authorities, especially in London where demand and costs have risen sharply, according to a new paper on arXiv. To address this challenge, researchers at the University of East London developed DOMUS, a cloud-based, AI-enabled decision-support system customised for the London Borough of Newham. The system integrates household case records, policy-constrained affordability and suitability rules, and live private-rental listings within a single governance-aligned workflow, aiming to standardise and accelerate placement decisions.
The DOMUS System Architecture
DOMUS combines transparent, rule-based filtering with large language model-assisted search, the paper explains. Household and property attributes are encoded into policy-consistent representations before AI-assisted ranking and explanation. The system applies bedroom need, affordability thresholds, geographic preferences, and accessibility requirements systematically, while preserving officer discretion and auditability. The platform is built as a modular, cloud-native Software-as-a-Service (SaaS) architecture, designed to be deployable across other UK boroughs and adaptable to other public administration tasks characterised by scarcity, rule-bound eligibility, and high stakes.
| Feature | Description |
|---|---|
| Rule-based filtering | Applies statutory policies on bedroom need, affordability, geographic preferences, and accessibility |
| LLM-assisted search | Uses large language models to augment ranking and explanation of suitable properties |
| Auditability | Preserves officer discretion while ensuring decisions are traceable and compliant |
| Cloud-native SaaS | Modular architecture designed for replication across boroughs and other public administration contexts |
The paper documents the creation of DOMUS from scratch at the University of East London, explicitly tailored to the needs of the London Borough of Newham. By encoding household and property attributes into policy-consistent representations, the system ensures that AI-assisted ranking aligns with statutory requirements.
Pilot Results and Operational Impact
A pilot deployment in Newham's secure environment evaluated operational performance relative to manual workflows. According to the paper, results indicated substantial reductions in search time, improved adherence to key placement constraints, and high staff satisfaction, while maintaining statutory compliance and role-based accountability. Although specific metrics are not detailed in the source, the qualitative outcomes demonstrate the feasibility of ethically governed AI deployment in local government.
The system's ability to standardise the application of rules such as bedroom need, affordability thresholds, geographic preferences, and accessibility requirements reduces manual effort and error. The preservation of officer discretion ensures that human judgment is not entirely replaced, but supported by AI-generated recommendations and explanations.
Scalability and Replicability
Beyond the Newham pilot, the paper frames DOMUS as replicable digital public infrastructure. Its modular, cloud-native SaaS architecture can be deployed across other UK boroughs facing similar pressures. The authors highlight that the system could be adapted to other public administration tasks characterised by scarcity, rule-bound eligibility, and high stakes. This positions DOMUS as a contribution to debates on AI-enabled public value creation in e-governance.
The development and pilot demonstrate the feasibility of scalable, ethically governed AI deployment in local government. The paper emphasizes that DOMUS integrates household case records, live rental listings, and policy constraints into a single workflow, enabling faster, more consistent placement decisions without compromising statutory obligations.
Implications for Technology Procurement Leaders
For enterprise technology buyers and digital transformation leaders in the public sector, DOMUS offers a concrete example of how AI and SaaS can address pressing fiscal and administrative challenges. The system's transparent rule-based filtering combined with large language models provides a blueprint for building decision-support tools that are both powerful and accountable. The modular architecture suggests that similar systems could be developed for other constrained resource allocation tasks, from social housing to emergency services.
According to the paper, the researchers designed DOMUS from scratch at the University of East London, customised for Newham's specific needs. This underscores the importance of tailoring AI solutions to local policy contexts and operational workflows. The pilot's success in improving search time, constraint adherence, and staff satisfaction provides evidence that AI can deliver tangible benefits in e-governance when implemented with appropriate safeguards.
As London boroughs and other local authorities grapple with rising temporary accommodation costs and demand, systems like DOMUS may offer a path toward more efficient and equitable placements. The paper's findings contribute to ongoing debates about AI's role in creating public value, demonstrating that scalable, ethically governed AI deployment is feasible in local government settings.
