Deflecting common questions with AI to improve customer support agent efficiency at MasterClass while maintaining CSAT.
MasterClass makes it possible for anyone to learn from the best. Students can access online classes from experts like Anna Wintour, Gordon Ramsay, Serena Williams in fields like fashion, film, business, and many more.
MasterClass is a fast-growing online education platform that creates expert-led courses around the world.
Due to their accelerated and continuous growth, MasterClass needed a tool or software to make their customer support experience more efficient and effective. MasterClass wanted to increase their customer satisfaction score while decreasing their first-reply time and total resolution time. To do so, MasterClass needed a way to deflect the Tier 1, ‘simple’, customer inquiry tickets so that the support agents can spend more time focusing on the Tier 2, ‘escalated’, support tickets.
Support agents were simply handling all tickets manually. However, the support team realized that Tier 1 tickets do not always need the attention of an agent. A ‘simple’ Tier 1 ticket, for example, could be something like a billing issue -- something like “How do I cancel my subscription” or “I want a refund”. Tier 2 tickets, on the other hand, are typically more technical and therefore commonly escalated issues. An example of a Tier 2 ticket would be something like “My video is too slow or doesn’t work.” These tickets frequently need a human customer support agent to provide the customer with a resolution.
With traditional automated answering systems in customer support, there is often a tradeoff between answering tickets quickly and customer experience and CSAT. However, with Agatha this challenge is eliminated. Using Agatha, MasterClass’ customers get accurate answers and agents don’t have to spend energy on mundane questions.
At over 90% accuracy, Agatha significantly outperforms other AI. Forethought’s Agatha was trained on MasterClass’ unique data to create dedicated models to support deflection. She uses Natural Language Understanding (NLU), an AI technology that understands the in's and out's of human language, to understand and deflect - and to ramp up qiuckly. Because she’s trained on MasterClass’ specific data, she has great models and can kick off her work supporting support agents in as little as 24 hours.
Agatha has had a measurable impact on the team. MasterClass’s customer satisfaction has increased by approximately 3%, while the first-reply time has gone from 2 to 3 hours to 15 - 20 minutes. Additionally, over 8% of support tickets are currently being deflected. This deflection is allowing the simple tickets to be dealt with quickly, and is allowing MasterClass’ Customer Support Agents to tackle the challenging sooner because they aren’t solving easy inquiries.
I think why we see this increase in customer satisfaction, is because now our agents are spending time with these more challenging tickets and responding back quicker because they don’t have to deal with these easy, kind of one-off tickets, and so that helps us [MasterClass] drive better performance overall.
In addition to bringing in Agatha Assist to support new agents, MasterClass is working with the Forethought team to make Agatha even more accurate. Specifically, MasterClass is focused on streamlining their internal processes to increase ticket deflection accuracy. This next step is where MasterClass and Forethought work together to enhance the customer experience.