Grofers is a low-price online supermarket that allows consumers to order products across categories like grocery & staples, beverages, personal care, home & kitchen, furnishing & home needs, household needs, breakfast & diary needs, baby care, pet care, meats & seafood and get them delivered to their doorstep. The delivery service is operational in 13 cities: Ahmedabad, Bengaluru, Chennai, Delhi, Gurgaon, Hyderabad, Jaipur, Kanpur, Kolkata, Lucknow, Noida, Mumbai and Pune.
Grofers was launching a Mega Sale in mid-January. They anticipated a massive spike in customer queries due to the huge marketing push behind the sale. Scaling up their customer support staff for a short period of time, to cater to the increased volume of queries, was not feasible. Automation was the only solution.
As a result, the Grofers team approached Haptik to automate their customer support. Their requirement – a chatbot that needed to be ready in less than 48 hours, in time for the launch of the Mega Sale!
Haptik built and deployed a Support Bot, integrated with the Grofers app, in just 2 days. This short turnaround time was possible largely due to two factors – the predefined templates of our bot-builder tool which enables swift bot development, and the proficiency of our team.
We adapted Grofers’ customer FAQ’s to a conversational format, and ensured a seamless technical integration of the bot to their platform. Our team worked closely with the Grofers’ team to resolve all their issues and queries in real-time, and to ensure the best possible experience for their customers. Once the bot was live, by examining the customer queries through the Haptik analytics dashboard, we were able to constantly re-train the bot to improve its performance.
The Top 3 customer queries that the Grofers support bot handled were About the Sale, Cashback/Refund Issues, and How to Redeem 100% Cashback .
The Grofers sale was live for 10 days. During that time, the bot was able to handle 75% of customer queries on its own, without the need for human intervention. For the remaining instances, where the bot was unable to respond to the customer query, one of the 6 human agents available resolved the query using the live agent-chat interface.