Results

• Nearly 285,000 transactions processed via automation over 8 months 
• 260,000 transactions processed without human interaction 
• 24,000 applications processed with human assistance (8% of cases) 
• 96% total success rate 

Background

Our client, a large UK government agency, was facing challenges with a transactional backlog that had grown from approximately 200,000 to 1.25 million. After brainstorming potential solutions, the client identified that Robotic Process Automation (RPA) would potentially help to automate these repetitive, low-level requests that their employees were currently doing manually.  

Davies was brought in to help deploy an RPA strategy, to collect scanned images prior to being sent to the client’s internal systems for processing and upload the relevant data into the same systems. In doing this, Davies and the client hoped to achieve a more efficient and streamlined way to manage paper transactions, reduce the chance of human error and enable existing teams to focus on higher-value activities. 

How we helped

Davies created a bespoke implementation of Azure Cognitive Services to read handwriting from the customer forms—this solution is outperforming many more expensive out-of-bag models. We introduced a Platform as a Service (PaaS), the UIPath Orchestrator. This now takes ownership over the problematic repetitive business processes.  

This was a new way of working and thinking for the client and required a solid channel of communication and support between Davies and the client, to ensure that the adoption of RPA was considered and correct for the client. To create an effective bespoke solution, we worked in a sensitive and agile fashion, continuously checking-in and adapting the solution to the client’s changing requirements.