Spark Energy is the leading energy supplier for tenants and let properties across the UK. Every month, thousands of Spark Energy customers call Spark’s call centre to make payments, submit meter readings, request altered bills and so on. This was causing a major headache for Spark Energy. Their call centre was regularly overwhelmed with calls at peak times in the month, causing service levels to drop and their review scores to fall below their aspired level. Motivation to answer repetitive calls would dwindle and the costs of managing these customers was rising in a highly competitive market. We challenged ourselves to make the customer experience better and lower Spark’s costs.
Innovation that started with measurable outcomes.
As with all proposed innovations at Equator, we start with a profitable outcome and work our way back. For Spark, this was reducing calls to the call centre and improving customer satisfaction. Both outcomes lead to revenue, directly and indirectly. We trawled through Spark’s detailed call centre records to discern the variety and repetition of calls and discerned that there were 5-6 types of calls that were highly repetitious but potentially fully automatable.
With this research, we were able to ascertain that, if we created an automated service assistant (bot) to handle these requests, it could potentially handle over 60% of inbound calls to Spark’s call centre. With wait times to speak to someone creeping up to 30 minutes and the call centre closed for more hours than it’s open, it was obvious we could deliver a solution that could drive real service change and savings for Spark.
Get that API feeling.
We knew that Spark needed a great, straightforward, no-nonsense chatbot that delivered on key tasks. It needed to handle new and exciting customer readings, monthly meter readings, online account signups, balance reviews and, critically, making payments. A good bot does key tasks clearly, quickly and simply. It doesn’t pretend to be your friend, it doesn’t do small talk and it doesn’t waste your time. Critically, it needs to do those tasks as well as, or better than a human, to be accepted.
To ensure this, we needed to make sure it was fully connected to Spark’s various systems to allow us to undertake those tasks quickly and correctly. Within 48 hours, our Innovation team had built all the necessary connectors to Spark’s systems, demonstrated the power and simplicity of the concept and convinced Spark this was the way forward. Fast forward to today and their bot “Ami” is available 24/7 to handle said requests.
No waiting around for results.
Spark Energy critically wanted to deliver the best service to their customer and they wanted to see that reflected in their customer service scoring. Thanks in part to the introduction of Ami, we have seen Spark’s “CustomerSure” customer service score rise to 7.9 and their Trustpilot move from a middling score to one that’s better than the big six energy suppliers. Alongside that, every request that Ami handles reduces handling costs to Spark by a considerable amount, allowing them to redeploy staff into outbound servicing and new business acquisition. A win-win scenario.