SPONSORED BY AWS
The Driver and Vehicle Standards Agency (DVSA) has used artificial intelligence to improve its vehicle-testing services, helping to make UK roads safer. The organisation has been on a seven-year journey to digitally transform the way it works, improving its services one by one.
The DVSA’s main role is to oversee MOT tests, testers and garages. Initially, it had been seeking a partner to replace a mainframe-based MOT system with a faster, more intuitive cloud-based alternative. It opted for Kainos, a digital services provider, and AWS Partner.
“The vast majority of our business over the last five years has been in cloud services, partly because our customers are demanding cloud solutions, but also because the solutions we can offer can be more flexible than any of the alternatives to the demands of the customers. Cloud can scale up and down more quickly and be more cost effective for customers,” says Peter Campbell, chief technology officer at Kainos Digital Services division.
Kainos Digital Services division has won a huge amount of business across the UK government’s G-Cloud and Digital Outcomes and Specialists frameworks. The former is aimed at easing procurement by public sector bodies of IT services that use cloud computing, while the latter is a framework for selling digital outcomes, specialists and research services.
Campbell explains that the inception of cloud, coupled with the service capabilities that cloud partners have, has levelled the playing field in terms of who public sector customers are procuring services from.
The inception of cloud has levelled the playing field in terms of public sector procurement
“The digital marketplace frameworks have been transformational for governments. Both allow government departments to buy against the market with a lot more transparency, but also [give] an opportunity for smaller businesses to market and sell cloud services. It has been a big part of Kainos’s ability to deliver and secure cloud work,” he says.
The new, faster, more intuitive MOT service is hosted on AWS and provides the foundation for DVSA to work on a number of other projects, including the Check MOT History and Get MOT Reminders service, which uses AWS technology to send test date reminders to motorists.
It was while Kainos was helping DVSA with the modernisation of IT infrastructure that the joint team spotted an opportunity to improve vehicle-testing services.
The DVSA carries out inspections to help ensure MOT tests on cars and motorcycles are conducted properly at garages. With 260 vehicle examiners carrying out unannounced spot-checks and 24,000 garages nationwide, effective targeting of garages is critical.
Kainos worked with the DVSA to create a machine-learning tool that continually identifies abnormal behavioural patterns in MOT testing and assigns risk ratings to individuals as well as garages. The algorithm builds a picture of performance over time, detecting garages with poor practices. The tool uses AWS to run applications and to securely hold large amounts of data, including test volumes, frequencies, pass rates and disciplinary histories.
The results have been remarkable. Better targeting meant the proportion of garage visits leading to an “unsatisfactory” or “needs improvement” verdict rose from 28 per cent in March 2018 to 77 per cent a year later. The proportion of visits resulting in disciplinary action or warning letters more than doubled.
Furthermore, the time taken by DVSA staff to complete pre-visit data analysis has halved and garages have welcomed the system’s more objective methodology and removal of rogue traders. Ultimately, the new vehicle-testing services tool is helping to make roads safer.
“Through our work with Kainos and moving the MOT testing service onto AWS, we have transformed the way DVSA detects poorly performing MOT garages,” says Alex Fiddes, DVSA’s head of digital operations. “Our priority is to help protect everyone from dangerous drivers and vehicles, and this technology is helping improve the standards of MOTs across Britain as we are better able to target dangerous practices.”
Cloud projects have enabled DVSA to keep improving its services. “We help our customers to develop their own skills and capabilities,” says Campbell. “We work closely with DVSA to build out their data science team, so they could start to innovate and do experiments in the cloud themselves. They’re using the machine-learning prototype as a template to work on their own machine-learning and artificial intelligence projects.”