Our client was a leading electronics product manufacturer with a presence in the Middle-East region, namely in the UAE, KSA, and Africa. With a catalog of over 4000 SKUs, the company mainly focused on selling these products via a network of distributors in the region.
The web-store allowed users to view the products online, but the shopping was routed via the distributor’s online store. In doing so, the brand received support queries related to inventory, new product releases, service centers, and product promotion and offers. Being a principal company, the customer service team had to engage with their consumers only via social channels as that was the main platform to receive queries.
Client:
JumboTimespan:
Jan’19 - June’20Key focus:
To educate people through AIService:
Artificial IntelligenceLive Preview:
https://www.jumbo.aeThe company’s contact center team had a limited bandwidth that could only handle a few hundred queries per day, limiting their ability to attend to queries credibly and in real-time. Thus, customer experience was taking a hit and so, the company decided to explore the possibility of using AI conversational agents to handle their customer support queries from across the UAE, KSA, and Africa.
The brand chose to work with Blue Logic Digital as their partner to develop this Chatbot.
The company briefed Blue Logic Digital to roll out a multi-lingual conversational AI bot to address customer queries in French, Arabic & English. Some of the key requirements for this interactive Chatbot were as follows:
Blue Logic Digital deployed a comprehensive tool – Google Dialogue Flow – to customize the entire conversational flow in a multi-lingual format, i.e. in French, Arabic, and English.
The key development technique used was entity extraction, which ensured that accurate and appropriate responses were given by the Chatbot. This entailed creating dynamic entities since product categories and sub-categories were fetched in JSON format.
The Blue Logic Digital AI team created base-level training datasets by identifying commonly asked questions. These datasets were then classified and mapped to a confidence score value. This helped in building the machine learning components to be able to track the intent and respond based on the cumulative scoring.
The solution involved undertaking the following tasks:
Engaging with Blue Logic Digital for the AI Chatbot solution rendered valuable results.
