Online food delivery is one of the most competitive markets where every penny counts and businesses spend millions of dollars on reimbursements annually. That’s why providing exceptional customer support is as crucial as ensuring smooth and safe refunding processes.
Building upon a previously developed chatbot, our team was tasked with upgrading its capabilities to address evolving security threats and optimize agent efficiency. This involved fortifying fraud prevention measures, streamlining agent workflows through strategic integrations, and unlocking the power of data-driven insights.
The goal is to have a single pin of glass for support agents, meaning one screen that they do their whole job out of rather than having to switch between apps. That’s a kind of our longterm strategy: whenever we can, if we learn of another internal corporate system that the client uses, we want to try and bring it into our experience so that agents become more efficient, they don’t ever have to leave the workspace. It’s a strategic thing for us and it’s also a huge efficiency game for the agents.
Our client, a prominent food delivery platform, faced a two-pronged challenge:
1. Escalating Fraudulent Activity: Their refund system was a target for bad actors aiming to exploit it for financial gain. These malicious attempts ranged from individuals seeking free meals to organized efforts to obtain and resell large quantities of food.
2. Agent Workspace Complexity: Support agents were bogged down by a disjointed workspace, requiring them to navigate between multiple systems to resolve customer issues. This constant toggling between applications hindered efficiency and increased the time required to process refunds.
To address the client’s challenges, we implemented a multi-faceted solution that leveraged our expertise in system integration and chatbot development.
Recognizing the critical need for a streamlined workflow, we created a unified agent workspace. This involved integrating not only the client’s CRM, loyalty program, and order management system, but also leveraging our connector development services to seamlessly connect a variety of other essential platforms into a single, cohesive interface. This eliminated the need for agents to switch between applications, providing them with a comprehensive view of each customer and empowering them to resolve issues quickly and efficiently.
Working closely with the client’s security team (and adapting to a mid-project change in security vendors!), we implemented robust measures to whitelist traffic and identify suspicious patterns. This included flagging potentially fraudulent accounts, blocking repeat offenders, and developing algorithms to detect anomalies in refund requests.
When the organization transitioned to a new security vendor, our team ensured a seamless migration of customized security features and swiftly adapted to the new vendor’s protocols. This ensured ongoing protection against evolving cyber threats.
We focused on removing friction points and enhancing clarity throughout the chatbot interaction. Users can now navigate the refund process effortlessly, with clear prompts, helpful guidance, and a conversational flow that feels natural and intuitive.
FAQ and Guided Flows:
To elevate the chatbot's self-service capabilities, we integrated an FAQ section linked to the knowledge library and implemented custom flows to guide users through common scenarios.
Real-time Integrations:
By leveraging APIs from DoorDash, Google Maps, and Apple Maps and many other services, we equipped the chatbot with real-time information to address specific customer inquiries.
Escalation to Live Agent:
For complex issues requiring human intervention, we ensured seamless escalation to a live agent, maintaining a consistent and personalized experience.
Custom Backend Recording:
We developed a system that feeds data directly into the client's business intelligence platform. This provided valuable insights into chatbot performance, user behavior, and emerging trends, enabling data-driven decisions to optimize the refund process.
We plan to leverage Generative AI and LLMs for intelligent routing, intent understanding, and agent assistance, further enhancing the efficiency and accuracy of the chatbot.
We are committed to continuously improving our fraud detection and prevention capabilities, including the development of more sophisticated algorithms to identify and block suspicious activity.
We continue to adapt to creative and challenging new rules devised by the client’s business team, as they continually evolve their logic for awarding automated refunds and offers.