Reducing 4 days’ work to a matter of minutes

Background

Described as “the United Kingdom’s innovation agency”, Innovate UK is a non-departmental public body designed to stimulate the economy by driving innovation in the marketplace. In 2017, the rising volume of applications led Innovate UK to consider a data solution to support their applications assessment team.

They required a solution that would enable them to quickly identify duplicate or potentially fraudulent submissions and that could assign applications to assessors with greater efficiency. In addition, the solution needed to be able to assign an appropriate assessor to review an application with greater efficiency than assessment and assignment by hand.

The solution

Cobalt Grant Manager utilises text analytics to automate the process of grant application checking and assessor matching, delivered via a comprehensive and responsive user interface. It automates the process of checking applications to ensure companies are not making fraudulent submissions.

The solution performs bulk document similarity checking for multiple submissions of the same application. Applications are scored against all previous applications, and a Natural Language Processing algorithm, developed by Analytics Engines compares sections within text documents.

Using Natural Language Processing and Text Analytics, the solution matches application topic areas with the skills and experience of an assessor to ensure a good fit.

Prior to deployment of the solution, manual assessment and assignment of applications took approximately 4 days to complete. Using Cobalt Grant Manager, the time taken to complete this process has been reduced from four days to just a matter of minutes.

The solution has also helped to improve the accuracy of the assessor matching process, resulting in a 5% reduction in the amount of rejected applications by assessors due to lack of skills match, compared with assessments by hand. At present, the solution is able to match assessors with an accuracy of between 96% and 97%.

The Analytics Engines team have been extremely communicative and supportive throughout the process and were very keen to find ways to assist and support us.

James Danek
Applications and Assessment Team Leader
Innovate UK