Challenges

Our client - a medium-sized hotel in New York - wanted to integrate an automated solution to its customer service department which processes a large number of requests on a daily basis.

It’s worth noting that the majority of those requests are mostly identical or related to a particular topic - e.g. reservations or background information. Therefore, the client decided to develop a chatbot application that would answer typical questions and reduce both the working time and costs of the support team. In order to provide the application with a natural feel close to a real support specialist, the application has been crafted using machine learning techniques which have drastically improved its “communication skills”.

Solution

The key feature of this chatbot is an API provided by Dialogflow and its webhook. Thus, our bot can learn via the list of created data or go outside Dialogflow to search for more information. The user graphic interface (GUI) is intuitive and comes in two versions: the first one is based on React Native for mobile, and the other one represents a web version built in React.js.

Backend part

Apart from integrating the APIs, our team has developed:

the MVP of this project; the list of data to our APIs on Dialogflow; the webhook that allows our chatbot to look for information outside Dialogflow;

Front-end part

- developed iOS, Android, and web applications;

Specific features

The feature that stands out from the list is the possibility to translate textual information in a variety of languages.

Results

The chatbot can be applied to a number of industries: traveling, healthcare, e-commerce, etc. Сustomers can get their answers in no time and find what they are looking for via the list of the chatbot’s data instead of waiting for the support specialist to do it.