TechPrufer
Get started
Grab
Grab logoGrab · Feature trail

Customer Support Workforce Routing

1 post · Feb 2021 – Jul 2021

What this trail covers

Grab has evolved its customer support operations by building an in-house workforce routing system. This system addresses limitations of third-party solutions by enabling granular priority management for requests, allowing for bespoke product customization to handle specific use cases like bulk configuration changes and dynamic queue limits, and providing enhanced reporting and analytics for agent productivity and performance. The solution uses separate queues for different priority levels and em

All Grab trails
  1. Reshaping Chat Support for Our UsersThis post details the product and design evolution of Grab's in-house chat support system. It highlights the MVP experiment that confirmed the hypothesis of a more stable native in-app chat, leading to a better user experience. Key scaling efforts include implementing dynamic queue limits based on Little's Law to reduce waiting times by ~30% and unresponsive users by ~7%, and introducing ML-based autocomplete suggestions to reduce average chat time by 12%. Further optimizations involved adding timers and nudges to reduce overall chat time by 22%. The post also emphasizes the importance of data-driven decisions, user testing, addressing root causes, and shadowing users.
  2. Customer Support Workforce RoutingThis post details the technical implementation of Grab's in-house customer support workforce routing system, focusing on Livechat. Key contributions include the design of a priority-based queueing system with time slices for de-queueing workers to ensure high-priority requests are handled promptly, and a mechanism to prevent request starvation. It also describes the approach to customising agent assignment with agent groups, dynamic queue limits, and proficiency checks, along with a locking mechanism for concurrency control and reallocation strategies for unaccepted or transferred chats. The system also enhances reporting and analytics for agent utilization and chat transfers.