Fetch’s Requirement to Solve for Rapid Product Expansion
Fetch is America’s most popular rewards app, working with brands across all categories to reward consumers every time, and everywhere, they shop. The app frequently adds new point-earning brands and personalized offers—up to 30 per day—that align with users’ shopping habits, enabling them to earn points on the products they love and discover new favorites, all while deepening their connection to the brands on and off the platform. The company has also been rapidly expanding the app’s core feature set by offering users the chance to earn points for playing games, eating at restaurants, shopping online, and trying out new apps, in addition to getting points for their groceries and health and beauty products. Additionally, the company aims to increase its US footprint from 1 in 10 households to 1 in 3 in the near future.
Every time new products or app features are launched, support tickets temporarily surge as customers navigate the changes. Eventually, this leads to a new fixed workload increase.
The combination of their app’s growth and a rapidly expanding user base meant they needed a solution to deflect some common, easily answered questions that their support team was getting repeatedly. But Fetch’s team has a healthy skepticism of chatbots — they’ve all been accustomed to lousy experiences with clunky keyword-based chatbots, and other software tools that were cumbersome to configure.
They needed a solution that wouldn’t be cumbersome to manage, would accurately understand user’s inquiries without making them jump through conversational hoops and would quickly provide highly accurate answers to their questions.
– Brian Johnson, VP of Support Operations, Fetch“The thing that really blew my mind, and ultimately, the biggest reason we went with Forethought, was the very creative approach they’ve taken to create workflows. The same three sentences I would type in Slack to tell someone how to close a ticket: that’s how you configure the bot. When I saw that, I felt like I was seeing the future. It’s one thing to see the bot do something but another to make it so easy to configure.”
Fetch’s Two Main Pain Points
Fetch currently employs more than 100 customer support professionals. They mainly communicate over email, interacting with thousands of users a day.
As they continue to grow, it’s critical they find a solution to accurately collect customer information before they speak with a live agent and to deflect basic user inquiries altogether.
1. They needed a tool to accurately answer fundamental, repeated customer questions
About 30% of support tickets are from customers who need help navigating the app or addressing questions that are easily answered in the help center.
To deflect these issues, Fetch initially used Forethought Triage to predict the content of a ticket. When possible, Fetch would send automated email responses based on that prediction, effectively deflecting many simple tickets. While this was a practical solution, it only allowed them to address less than 1/3 of these types of tickets.
2. They needed a tool to gather information from a customer before talking with a live agent
Fetch found that an additional 20% of their tickets were inquiries that required support agents to gather additional information from the user in order to clarify and address their issues. This created an alarmingly high reopen rate across the entire team.
A Rigorous Evaluation of Four Market Leaders
The Fetch team was incredibly skeptical of automated customer support. Their negative experience with chatbots and other enterprise CX software platforms prompted them to rigorously test potential solutions before making a final selection.
Their evaluation criteria included the following:
- The tool needed to be implemented with minimal engineering support
- They didn’t want any software development kits (SDKs) inside their app
- Their team required full integration with Zendesk
- User information had to be encrypted throughout the whole journey
- They didn’t want users to have to repeat information at any time
- The tools needed to be able to learn quickly with minimal configuration
- They wanted to be able to provide product feedback to the vendor influence the product roadmap
They evaluated four market leaders. During the initial demonstrations, it became clear that Forethought and one other vendor had a massive advantage in terms of ease of implementation and optimization.
“That first Forethought demo was what I describe as a major AHA moment. It was so far beyond what I thought automated support could do that it became apparent that this direction would yield a lot of value for us.”
They ultimately only tested Forethought and one other vendor to see whether their automated agents could effectively use existing knowledge base content to answer 35 common ticket types.
The team then manually reviewed results to determine if correct answers were given and sufficient customer data was gathered.
Forethought was selected after testing 93% accurate—a full 28% more accurate than the other tested competitor.
Forethought Solve Chosen for 93% Accuracy and Incredibly Easy, LLM-powered Configuration
The Fetch team was impressed by Forethought’s accuracy score and their past experience with Forethought Triage. In just 11 months, they automated 316,000 ticket actions for $90,000 at a 3.9x ROI.
They ultimately chose to deploy Forethought Solve because of its accuracy and full LLM-based configuration interface.
“Since deploying our Forethought AI agent, Scout, our users have been able to get a wide range of questions answered. The CSAT scores for fully automated chats are as good or better than those with human agents, due to the speed at which their questions are addressed. Additionally, the breakthrough innovations Forethought has added to their platform since we launched has opened the door for us to massively expand the scope of issues that Scout can handle with and without human assistance. While we were excited going into this project, since launching it has become clear that we vastly underestimated the potential impact of this transformative technology.” Brian Johnson
At Forethought, we believe the future of customer experience is generative AI. Forethought is the most advanced generative AI agent for customer support, trained on proprietary data to automatically resolve more issues at a lower cost and elevate customer satisfaction with every interaction. Request a demo to see how we can improve the accuracy of your automated support.
