Our newest journey teaches you how to create a banking chatbot with Watson™ and Node.js. There’s no doubt that chatbots are ubiquitous. They have become our personal assistants and so much more. No call queue should ever be needed to get through the usual support conversation on any website with cognitive chatbots. Market analysis suggests that the financial sector in India is betting big on chatbots – which is one reason we have created a journey that is a banking chatbot specific to India. Globally, banking bots have been conversing with clients beyond traditional transactional queries for some time now – a service that Indian banks have only just begun to dabble in because of the regulatory norms put in place. This, combined with the power to transform an industry, is one of the many reasons why our focus on banking chatbots in India is particularly exciting. The cognitive banking chatbot we have created uses many Watson technologies, including Watson Conversation. For those unfamiliar, Watson Conversation allows you to quickly build, test, and deploy a bot or virtual agent across mobile devices, messaging platforms like Slack or even on a physical robot (we use a web interface here, but of course it can be easily adapted). Watson Conversation has a visual dialog builder to help you create natural conversations between your apps and users, without any coding experience required. The chatbot also uses the Watson Discovery passage-retrieval feature, which makes it easy to extend a chatbot with a collection of FAQ documents. Without any document conversion or discovery training, this shows what you can do by just dragging and dropping Microsoft Word docs and adding a simple call in your JavaScript. Admittedly, you should probably do training – but it works as if by magic. In the “Create a cognitive banking chatbot” journey, you can see how to make a Node.js chatbot that can pull answers out of a collection of documents. While we only highlighted two of the cool technologies, you’ll notice that if you follow along with this journey, it is packed with other cool features that can be done with just a little code: * Add emotion detection so your chatbot can mimic and appropriately respond to the end user’s sentiment 🙂 🙁 😐 * Add natural language understanding to recognize an entity, such as a location * Have your chatbot reply include buttons for users to click The banking conversation also includes many good examples of how to interact with a user and back-end systems, which of course you can modify to your liking (or not). Get started with our banking chatbot journey today. You’ll find code, step-by-step instructions, architecture diagrams, and more.

3 comments on"Create an FAQ-aware cognitive banking chatbot with Node.js and Watson"

  1. Raymond_Buckland August 17, 2017

    Link to ‘journey’ is broken.

  2. Neil MacKinnon August 22, 2017

    Thanks, Raymond! Links are working now.

  3. Raymond_Buckland August 23, 2017

    Neil, thanks a lot… all working. Much appreciated.

Join The Discussion

Your email address will not be published. Required fields are marked *