AI has rapidly evolved from a mere amusement to a fundamental tool that empowers online users to take care of complex activities without the need of a live agent. Conversational AI solutions offer the ability to create high-quality chat bots through which customers can schedule appointments, check order status, answer product questions, and much more — without pressure on the call center.

In a previous post, “Building a Reservation Bot with the NLX Conversational AI Platform”, we guided you through the steps of building such a bot tailored for a reservation service. Today, we demonstrate how banking support can be redefined with the implementation of a real-time bot built using

Customers of the fictitious “bankRTC” can chat with the bot to accomplish simple tasks like checking an account balance and even more complicated ones like paying a specific bill and setting up recurring payments. If additional support is needed, we’ve added the option to communicate with a live agent via a video or audio call powered by the Vonage Video API. All within the same web interface!

Tech Stack 

We used the following tech stack for creating the banking bot:

  • Chat SDK for NLX Bots – The official JavaScript SDK to communicate with conversational bots created using NLX Dialog Studio.
  • Material UI – Providing pre-built React components that ensures a sleek and consistent design throughout the application.
  • Next.js – An excellent web development framework that leverages the benefits of React dynamic components plus the improved performance and SEO of server-side rendering.
  • Vonage’s OpenTok.js – The official library for video sessions powered by Vonage Video API in web applications.

Building a Real-Time Chat Bot 

In order to develop this bot, not only did we need to build out the conversation flow, we also needed to authenticate the user and fetch data from the API to show user’s account details. It was an interesting challenge, especially using data in the conversation nodes.

With Data Requests nodes, we are able to fetch real-time data from the API and handle request failure and success scenarios. The “Loop” node allows us to traverse through lists of data and then use and display these accordingly.

Integrating Audio and Video Communication with a Live Agent

For maximum flexibility in the customer-agent support conversation, we integrated the option to connect to a video or audio call with a live agent.

We connect to a Vonage OpenTok session using the NLX Chat SDK via an access key and token. Once connected to the session, we create a Publisher object to publish the user audio and video, and we also create a Subscriber to receive media streams from the agent that’s taking the call.

Time for a Demo!

Redefine Customer Support for Your Organization

This simple demonstration provides a glimpse into the many possibilities that AI can offer to improve customer support experience not only for banking but also many other industries. NLX provides a straightforward approach to build powerful bots without writing a line of code. Vonage Video API allows to add video communication capabilities to web applications. Next.js and Chat SDK for NLX Bots allows developers to create web applications that interact with NLX Bots.

If you’re looking to improve your customer support experience and leverage the benefits of chat bots and other AI-enabled workflows in your real-time communication application, has the expertise you need! Contact us today and Let’s Make it Live!

Recent Blog Posts