Grab Customer Support agents (CE agents) handle hundreds of calls per day. Unfortunately, the information they need to answer these calls are spread over different data sources. Compounded with the volume of calls they handle, this leads to an increase in Average Handling Time (AHT), which is approximately 3 times that of the company's Service Level Agreement (SLA). Thus, reducing AHT helps CE agents be more productive in their day-to-day work and lead to costs savings for the company.
Project Phoenix aims to optimise Grab's customer support agents' workflow by providing all necessary information into one single platform. Agents will be able to locate pertinent information across different databases and efficiently triage the most common calls they receive.
The pain points Project Phoenix solves are:
1) Lengthy Average Handling Time
2) Increasing backlog of calls and reduced user productivity
3) Use of multiple applications and open web browser tabs to answer one call
Prior to designing Project Phoenix, multiple design workshops were planned and conducted, where users and stakeholders were invited to partake in card sorting, affinity mapping, as-is and to-be mapping, and ideation, to name a few. During the design process, user testing was conducted regularly at intervals, and further iterations were made based on the results gathered.
Password to the full Invision prototype will be available upon request.
I conducted design workshops in the initial stages of the design process with the users, the CE agents. Pictured here is a card sorting exercise, with the objective of determining what information is required in the new Phoenix platform. We started with information from the old platform on the cards, and had the users do affinity mapping, allowing them to group them together, or deem the information unnecessary and remove the cards. In doing so, I understood how the users viewed the information, and in my design, I displayed the information in the same groups together. I believe in letting my users inform my design decisions, as I believe this is key to crafting an experience that will fit their needs most.
This is the driver page, one of the pages the CE agents will work on most, as their job mainly consists of addressing calls and issues from drivers and passengers. Previously, to address an issue from drivers, the CE agents needed to access several different screens and platforms as the necessary information aren't consolidated, meaning that it takes a lot of time to gather information and address an issue, increasing AHT. Thus, I consolidated all the necessary information to address calls in the driver and passenger pages, to reduce the number of screens needed to addresses issues, in doing so aiming to reduce AHT.
The information that goes into this screen is determined by two ways: firstly, by card sorting and allowing the users to direct what is necessary, and secondly, by shadowing the CE agents in their daily work, observing the tools the use, the information they access, and the calls and issues they receive.
Information in this screen is ranked by importance. As the CE agents need to handle calls on bookings the most, bookings are placed higher. Similarly, as the second most number of calls the CE agents receive for drivers are on incentives, information on incentives are ranked second.
I also worked closely with the content team and the CE agents to determine how to word certain phrases to best aid them in understanding the words refer to.
I wrote research plans and conducted user testing sessions with the users, allowing my designs to be informed by the findings and iterating from there. During the project, I conducted a series of user testing sessions, each building on the insights gained from previous sessions. Pictured here is a snippet of the research plan.
I modelled this scenario after one of the most frequent questions received from the drivers, so as to test if the users were able to complete one of their most frequently executed tasks on the new design.
This is the Booking Card. Similarly, decision on what information to display is determined through card sorting exercises with the CE agents. As the action item of many of the calls is for the CE agents to "Cancel Booking", the button is conveniently placed at the top right. Other information ranked most important by the CE agents were also placed at the top, like booking code, status and payment methods.
One pain point the CE agents had was that it was hard to location information like pick-up address and booking time, as everything was not sorted. Through affinity mapping with the users, I grouped the relevant information together so it was easier for the CE agents to locate them.
In a similar concept to the driver page, I placed all the information a CE agent needs to address a booking issue in this booking page, aiming to solve the pain point of having to access multiple screens in order to solve an issue.
This is the incentive page, where all information necessary to solve issues on incentives lie.
Prior to this design, one of the pain points CE agents faced was that they could not keep track of whether drivers were eligible for a specific incentive, as it was troublesome to compare the driver's rating, Acceptance Rating (AR) and Cancellation Rating (CR) against the respective minimum eligibility ratings the incentive imposes. At the same time, drivers often debated with CE agents on how much incentives they should earn. The progress bar thus helps give a clear view on how much incentives the drivers are actually qualified for.
This is what the Driver Card will look like if the driver has been blacklisted. There are certain use cases where the driver has to be blacklisted, like if the driver has repeated committed fraud. Once blacklisted, the driver can never drive for Grab again. Thus, I greyed out the driver page, indicating that this driver has been blacklisted, in doing so indicating that actions cannot be made on this page anymore.
This is the passenger page, which is designed in a similar concept as the driver page. Information here is also gathered through card sorting with the users, as well as observing their their as-is flow by shadowing them in performing their daily tasks.
Again, information displayed is ranked, with most important (ie most number of calls regarding bookings) at the top. Unlike the calls from drivers, most passengers tend to call about promo codes, after issues regarding their bookings, which is why appeasement history is placed just below bookings.