The setup of a successful AI Chatbot
How to setup a successful AI Chatbot for your clients? Let’s go through some of the setups we have done for our clients in the past few years. Below are some of the real-world examples of how we setup our clients’ AI Chatbot.
Fast and easy update
Once we have the AI Chatbot up, running and trained, most customers would probably just want to update the responses of their AI chatbot fast. For those basic questions and answers type of conversations, we use Microsoft Excel and upload them into Microsoft QnA Maker.
The chatbot trainer can then easily reupload or replace the excel files. It is easy enough to do it via the portal of Microsoft QNA Maker, but we simplify it via XTOPIA. Here is a quick tip: For short-lived QNA like a campaign or contest, it is recommended to store them into separate Excel files so that it can be easily removed. You can upload many different copies of Excel files into Microsoft QNA Maker.
What if you have dialogs that are not using Microsoft QNA Maker? With XTOPIA’s Chatbot Apps, our users can also update the responses of the bot easily. Once setup, they can login to XTOPIA and look for the Dialog Manager and update the dialog responses or questions. The responses are also multi-lingual as well. We use Microsoft Text Translator to translate user’s input. You can also choose other text translators such as Google Translator since the bot framework is very flexibility.
What about data that is in tabular format such as forex exchange or the latest price of petrol? You can store the data in XTOPIA’s data source and connect the AI Chatbot to the data source via XTOPIA’s workflow. Then, the user just needs to edit the data source or import an excel file into the data source.
An AI Chatbot can do more if you setup the chatbot to interact with your website. It does not need to be just a small window embedded in the corner of your website and do simple conversation. Let’s talk about what we can do when we integrate the website with XTOPIA. One example is the page redirection and another one would be the page element manipulation.
In the example below from InvestKL website, Kayla can redirect users to the Contact Us page for them to fill up an enquiry.
In this Sype’s example below, the chatbot is updating the page’s dropdown as user choose the options on the fly. You can give it a try on the website.
Another setup to consider is to choose a platform that can help to save cost and scale as needed. We chose the Azure Bot Framework which is hosted on the Microsoft Azure’s App Service. App Service can be scaled easily and can be very cost effective. When the bot has very low usage on normal days, we can keep the cost low by using the lower plan of the app service. If the chatbot is popular during sales season for example, we can scale it up easily and only pay more for the peak seasons.
We hope to find more ways to make AI Chatbot more useful for our customers in the future.
XIMNET is a digital solutions provider with two decades of track records specialising in web application development, AI Chatbot and system integration.
XIMNET is launching a brand new way of building AI Chatbot with XYAN. Get in touch with us to find out more.