Wilson Parking | IVR Proof of Concept
Introduced Conversational AI technologies to optimise the call centre processes. VERSA was engaged by Wilson Parking Australia to expand upon their current call centre technology (Amazon Connect) and optimise their customer service offering.
Wilson Parking Australia is responsible for managing hundreds of customer enquiries each day through a dedicated national customer service centre.
As industry leaders, VERSA was engaged to determine how Conversational Artificial Intelligence (CAI) technology could be introduced into the existing Interactive Voice Response (IVR) system built in Amazon Connect. Our goal was to add immediate value through a proof of concept that could alleviate the strain on their call centre, while providing crucial insight into the segmentation of caller queries based on existing service categories.
Our team undertook a Pilot project, delivered using a lean approach to agile delivery.
Our tried and tested processes lead to critical discoveries
Through our discovery workshops, we learned that an estimated 40% of enquiries related to refunds - an task that could be completed online. This insight formed the basis of our proof of concept, a CAI solution that could accurately understand and direct caller enquiries while guiding suitable refund use cases toward self service channels.
Our team started with the modeling of a proposed Voice User Interface (VUI) diagram. This comprehensive flow diagram mapped the skeleton of our new solution design against the existing call centre flows on Amazon Connect.
Through our process, we developed the ability to analyse detailed customer insights, exposing opportunities for optimisation and automation. We implemented an agent whisper flow in Amazon Connect to help reduce handling time and prepare agents for incoming calls.
Diverted an additional 6,000 calls per month.
The pilot solution saved over 46 hours worth of potential handling time. This equated to 2,000 fewer calls being handled by the call centre, when compared to volumes forecasted from year-on-year trends.
By the end of the training period, the AI was able to understand, categorise and service 71% of caller cases. The other 29% were captured through the fallback intent, then passed onto an IVR system to further progress these enquiries.
We able to provide Wilson Parking with a clear roadmap of optimisations that included a number of new intent categories and self service opportunities, giving them the potential to divert an additional 6,000 calls per month.