GRAB

Grab is a Southeast Asia's leading transport, logistics, and payments platform that operates in eight countries across the region. Grab has grown to 45 million downloads and up to 2.5 million daily rides.

PROBLEM

The current Grab's Help Center is not very efficient for CE Agents to handle calls and tickets. I was looking for eradicate the pain points of CE Agents and reduce the volume of enquiries.

METHODS

Observing users
Kickoff workshops
User testing

Getting started with empathy

Before I jumped into creating new solutions I spent a day with our internal CE Agents to see the world through their eyes. The time spend with users always reveals a tremendous amount about the way they think and the values they hold. Once I better understand users I conduct kickoff workshops and invite Business Owner, Project Manager, CE Agents, and Engineers to join the meeting.

Generating Design Ideas

After going through user journeys and stories it's time to toss around ideas that cover pain-points and opportunities we've uncovered. During the ideation process, we focus on generating as many ideas as possible. In this case, the quantity is more important than quality. Two main ideas egressed: either we hire more people or invest in developing AI/NLP. Business Owner who has the final call in the room decided that we should invest in the future and focus on the chatbot function. 

Prototype

Once we've determined what to prototype, it's time for the fun part and builds it. I split the group into 3 teams that have drawn wireframes for the most common complains from users. They consist of: passengers who lost an item in a taxi, passengers want to change a phone number and finally, passengers want to check the delivery status.
 

Test

Instead of conducting a typical usability testing with 5 users, we've decided to roll out the beta version to 5% of users. The reason being is we want to collect as much data as possible to improve the chatbot. My designs emphasize that the feature is only a beta version and I hope that it won't influence the retention rate. In this case, the time is the key and we make sure that we improve chatbot every day. 
 

What's next?

In the future, we'd like to handle not only the most common complaints but also those that are rare and unique. The current function is available only for Indonesian users. Hopefully, we will improve the MVP product soon enough to release it to Singaporean Passengers. What's more, I consider conducting another session to uncover Drivers most frequent complaints. My ultimate goal is to make the chatbot smart enough to replace the entire Help Center and make our users lives much easier.