Objectives of the service
Through this project UrbanTide is creating a new service and AI module called uSmart REVENUE which is utilising AI and Machine Learning to automatically spot fraud and identify missed revenue by integrating open datasets (examples include; Companies House, Charity Regulator, Food Standards Agency) with closed datasets (Local Authorities own internal Non-Domestic Rates valuation roll data) with Earth Observation data and satellite imagery, enabling us to label specific businesses across a wide geographic area.
Users and their needs
The cost of fraud to local government is estimated at £2.2 billion annually. By automating fraud detection, uReveal could help generate up to £350,000 a year in lost revenue for each local authority across the UK. This is money that local authorities can use to fund services and make communities better places to live.
uSmart Revenue is an AI-powered data insights platform that uses new innovations to identify and recover lost revenue from business fraud.
Service/ system concept
The tool does this by integrating a variety of open and closed datasets and using AI and machine learning it can automatically spot fraud and identify missed revenue.
By tracking the differences over time, local authorities can automatically detect changes, and therefore where businesses are due to pay a higher rate, without the need for manual input.
uSmart Revenue is a cloud-based software-as-a-service (SaaS) solution that users access via a secure web interface. Local authority non-domestic rates valuation roll data is uploaded into uRevenue, and the system handles the rest:
-
Data is geographically linked with a spatial location on a map
-
Satellite imagery from that area is used when available (or a new capture ordered) when not
-
Buildings are detected using AI algorithms and linked with the valuation roll data
-
Two images with a different time stamp are compared to highlight differences in building outlines and other rateable elements
-
The floor area of premises is used to evaluate any potential increased floor area
-
The imagery and public sector data is linked with OS mapping and graphically presented in the uRevenue tooling interfaces
The technical aspects of the above approach are illustrated in the following architecture
Results can be provided in a series of ways, including:
-
Fed into their existing management system via an API;
-
As an interactive map-based interface that can be integrated into existing BI tools;
-
As a raw dataset output for manual integration and analysis; and
-
As a set of concise reports localised for their area.
Space Added Value
We are building uRevenue utilising 0.30m and 0.50m satellite imagery from market leading providers who are able to task satellite capture over areas as required. The AI enables assessments to be undertaken at scale and faster than any manual approach often used by competitors. We are also exploring the value of the latest thermal imaging satellites to help support identify of building usage as part of the valuation process.
Current Status
uRevenue has passed BDR/ CDR and is well on its way to creating a usable and testable service within the pilot phase. It has secured access to Analyse Local and to the IRRV for supporting in this process and is on plan to create a new valuable product to support the industry.
Prime Contractor(s)
