Since mid 2017 Spark has reduced its customer service costs by 25 per cent and call volumes to its contact centre by 33 per cent through the use of human online chat and virtual assistants – chatbots.
The figures were revealed by Daniel Cooper, head of care, transformation and premium care at Spark, and John Wood, virtual assistant product owner, in a keynote presentation at Verint’s customer conference, Verint Engage, in Sydney earlier this month.
Cooper said that, in mid 2017 Spark had 1350 team members in its contact centre, “That was 23 per cent of the workforce dedicated to answering the phones or chat,” he said.
Voice was still the dominant channel for customer care, accounting for 88 per cent of interactions and chat was flatlining at 12 per cent of interactions, despite having been running for three years.
Since then, he said Spark had managed to get 70 per cent of customer contacts onto digital channels and achieved a 20 per cent reduction in contact centre work effort.
Call volumes are down 33 per cent since July 2017 and interaction volumes down 21 per cent, Cooper said, despite a nine percent increase in customer numbers in that time. He said customer servicing costs had been cut 25 per cent year on year.
Chat has risen from 12 per cent of interactions to 30 per cent. Spark’s chatbot, based on Verint’s Virtual Assistant, accounts for about 18 per cent of these interactions and is able to resolve about half. Cooper said he expected this figure to increase as it as Spark’s implementation of Virtual Assistant – dubbed Ivy following a call to staff to suggest a name — was integrated with Spark’s CRM and billing systems.
“We are reasonably happy. We have a lot more work to do,” he said. “The journey will continue. Our customers and shareholders demand more value out of us. And the way we deliver that is to make customer journeys more frictionless.”
According to Verint its Virtual Assistant uses artificial intelligence and machine learning to create a modern engagement experience. “Our conversational AI can predict what a customer wants, personalise the interaction, and determine the best next action to deliver a successful outcome,” it says.
Wood said the initial use of Virtual Assistant had been for Spark’s video on demand service Lightbox and after initial teething troubles, its successful implementation for that business had led to be being adopted across the main Spark business, replacing the incumbent system.
Cooper said initial difficulties with chat has resulted in strong internal resistance to its use. “There will be internal pressures, calls to turn it off. You have to manage those and leave people to get on with it.”