Mitgo tests ChatGPT for automated first-line support with a more “human” approach
With Mitgo and its businesses actively growing in India, the APAC region and Southern and Northern America, and platforms like ConvertSocial generating lots of user queries, the need for accurate and attentive first-line customer support is as important now more than ever.
As it stands, Mitgo’s first line of support receives thousands of requests through our various different channels. Of these issues, 80% are resolved and closed there and then, while the remaining 20% – usually more specific, advertiser-related issues – are forwarded to our business teams.
This all requires resources. Traditionally, this meant manpower. Can ChatGPT’s AI technology step up and offer our users the high-quality help and support they deserve?
Machine learning with a human touch
Our goal is to provide automated-yet-“humanised” technical support. A bot is currently being trained to learn the ins and outs of our company and how to answer some of our most-asked questions.
Working alongside a “live” technical support person, the bot will give us the chance to answer more questions faster and in a wider range of languages. It may also be able to take up some of our support team’s heavy lifting.
According to Zendesk statistics, 40% of customer support requests are resolved with one response (one-touch tickets).
In a classic example of this, a client asks a question, the answer to which can be found in our company documentation. A support agent has the choice of either referring the client to the documentation itself (which would involve the client finding the answer themselves), or writing a summarised answer to the question.
From a user-focused point of view, this second option offers far better customer service. But it also takes more time and manpower.
ChatGPT, on the other hand, would excel at this task – quickly and concisely giving a precise answer and relevant summary of the documentation.
If our AI bot could solve these 40% of queries, important human resources could be freed up to focus on more complicated cases.
Our AI switchboard operator
Another important aspect that our AI bot can help us with is the classification, ranking and forwarding of incoming requests.
By analysing the content of each question or query, our AI bot can cluster similar questions together, prioritising the most urgent and forwarding them to the appropriate teams.
Working essentially as a secretary or switchboard operator, our AI bot could help us improve our response times – dramatically lowering the time our clients wait for an answer from our current average of 3 hours.
Working hand-in-hand with technology
Should it prove to be a success, the results of our experiments with AI will filter down through all of Mitgo’s businesses, improving the already-high level of transparency and customer service we offer wherever it can.
Stay tuned for more updates and news coming your way soon!