The Next Generation of Customer Support Systems

Introduction

For quite sometime, I have been working in the domain of designing customer support systems for B2B companies. I honestly believe that effective customer support is the backbone of any good business today.

For past few weeks, we have been discussing in our team about new ideas for customer support and how can we take it to the next level. Recently, I got to know about Deepgram from the hackathon on DEV. And It clicked me immediately that how we can leverage it to build the next generation of customer support systems.

My Deepgram Use-Case

Imagine that you use a software for doing some critical work everyday. Now one day you encounter a problem and the software starts to misbehave. After some unsuccessful attempts to try and fix it, you decide to raise a ticket on their support portal. When you go the support page, It presents you with a big form which asks for so many details about your problem.

The above mentioned scenario is pretty common and almost all the companies have a similar support page for their customers. Now it may sound reasonable to fill up the details and raise a ticket but this is not the ideal user experience for a customer who is already irritated because of a not-working software. Also, written details will not provide the sentiment of the customer to the companies as well.

What if the customer can just speak about the problem he is facing instead of manually filling up a form? What if the companies can analyse the sentiment in customer's tone and decide the priority of the issue based on that?

Dive into Details

The system would utilise Deepgram's speech to text feature. Customers will need to file their complaints in the form of audio clips. The support portal itself can have a record button which customers can click and start explaining their problem. The backend will convert speech to text and analyse the problem to create a support ticket. We can have AI for extracting relevant details from the transcript of the audio and create a support ticket for customer.

Another feature will be to do sentiment analysis based on the customer tone, statistics on the time taken to explain and it will be used to determine the priority of the ticket.

We can also provide this feature only to the premium support customers so that they do not need to spend time filling up forms for frequent usecases.

Conclusion

Seeing the documentation of deepgram, I can definitely say that it's easier to integrate with it compared to other platform. This hackathon gave me good amount of exposure to the APIs and features provided by Deepgram. We are definitely going to try this on our customer support portal as a pilot program and we will monitor how customer responds to it.
 
Nonetheless, evolution is a continuous process and we will be experimenting with new ways to explore more dimensions of Customer Support.
 
Thanks for giving it a read guys. Cheers 🍻