HOOPOE our Customer Service Chat BOT is built and geared towards the banking industry targeting banks that deal with a lot of customers and have large sets of FAQs. The aim of the BOT is to facilitate service and products inquiry, minimize the interaction between customer service agents & customer and enhance the user experience

The main goals of the solution are:
• Reduce the number of customer service agents being utilized by the organization
• Improve the overall customer experience by offering new services
• Gain more insights on customer service performance and experience through the integrated analytics components.

Utilizing the latest technologies to insure optimum results for your customers, HOOPOE is a Machine Learning based solution that is hosted on the cloud infrastructure. With support for multiple languages as well as a diverse range of platforms that can be easily integrated with enterprise solutions.

Through its Advanced response system, Campaign integration, Advanced Analytics and 3rd Party Integration, HOOPOE guarantees an improved and more efficient customer service experience.

Cognitive/Artificial Intelligence Module
Hoopoe’s AI Module adds a new layer of data enrichment using propriety components developed by Hyve. The AI module uses advanced machine learning techniques to enrich the data and add new layers of information that enables the bot to understand the conversations related to personal banking In English, Arabic, Egyptian Dialect, and Franco Arab.

Sentiment Analysis
The first layer of data enrichment is AI based sentiment analysis. Through several training cycles and continuously improving algorithms Hoopoe provides sentiment information for customer messages. Hoopoe supports several languages such as Arabic, English and Franco Arab. The bot can automatically flag offensive messages and even execute a predefined action based on it, like warn the user, end the conversation, or transfer the chat to a live chat system.

Entity Extraction
Hoopoe AI provides an entity recognition engine that allows the system to detect and tag all posts containing important entities. As an added layer of enrichment. Entities are important to understand the conversation. The admin doesn’t need to define the exact entities in the system, just to add some examples and the AI engine will predict most the other forms of the entity which will dramatically increase the bot vocabulary and accuracy in detecting entities.

Intent Analysis
The Intent Analysis is based on a deep learning model to insure 85% accuracy level of Intent classification. The intent analysis component supports Arabic, English, Arabic Dialects and Franco Arab. The AI Engine already trained on personal banking intents using hundreds of thousands of banking questions.

Hoopoe

by Hyve for Technology Consulting