Case Study: Bilt
"Working with Decagon has been nothing short of phenomenal. The team has taken our extremely complicated data and created a tool that allows our customers to seamlessly receive help across our business. Additionally, the team has worked tirelessly to build integrations into our CRM and support setups to make sure we would not have to change any of our current operational workflows. Couldn't recommend them more highly!"
Thatcher Foster
VP, Client Solutions at Bilt
Bilt Rewards is a technology company that allows renters to earn points on rent and create a path towards homeownership. Bilt works with some of the country's largest multifamily owners and operators to build the future of loyalty programs for property renters. As of October 2022, the company's loyalty program and payment platform was rolled out to more than 2.5 million apartment units across the country. Users earn points and improve their credit by simply paying rent each month. Bilt's points can be used in 12 loyalty programs, including major airlines, hotels, travel, fitness classes, purchases, credit toward rent or a future downpayment. By mid-2022, Bilt had already processed over $3.5 billion in annualized rent payments and over $1.6 billion in annualized card spend.

As a consumer-facing product, Bilt has a large customer support team that handles volumes of customer requests each day. Moreover, as a product that deals with customer financials, getting customer support right is not only essential, but also a big competitive advantage. With recent advances in AI, Bilt wanted to explore how they could leverage AI to improve their customer support experience.

High-quality, fast, and personalized customer support is a critical component of Bilt's strategy. Well-executed automation unlocks higher customer satisfaction as well as massive reductions in cost.
The Problem
Traditional customer support automation wasn't good enough.

Before implementing Decagon, Bilt had been using a popular incumbent solution to deflect customer support tickets. The primary issue was twofold. First, the level of complexity of the business logic and data involved in resolving each customer support ticket was too high for the team to fully handle with previous-generation solutions. This led to a large overhead for the team to maintain decision trees, canned responses, and other rules-based systems. As the product evolved and things changed, this compounded into more and more manual work for the team to maintain.

Second, the customer experience with the incumbent solution was painful. Due to the integrations required into internal systems, many simple tickets could not be resolved automatically as the bot didn't have the ability to extract the right data and make the right decision. Furthermore, canned responses resulted in the same poor customer experience that many other bots provide. Responses were not personalized and often did not answer the customer's question, which in turn, led to more tickets and more work for the team.

For the team internally, the lack of high-quality tools to gain visibility into the conversations every month led to the support operation being more of a cost center than a growth driver.
The Solution
Decagon is built from the ground up with the latest advancements in generative AI. Large Language Models are used not only to draft a personalized answer, but also to handle the business logic needed to fully resolve a ticket, including calling various internal API endpoints to fetch relevant real-time data.

Furthermore, Decagon's powerful admin portal enabled the team to have complete visibility into every conversation, high-level metrics such as CSAT, and detailed insights. For instance, every conversation is automatically tagged with the topic of the ticket, so the team can easily see what types of tickets are coming in and how they are being resolved. This enables the team to make data-driven decisions to improve the customer experience and reduce the cost of the support team.

Importantly, Decagon and Bilt worked closely to integrate Decagon's AI into Bilt's existing workflows. This included integrating into Bilt's CRM, Zendesk, and other internal systems. This also included teaching the AI about the internal APIs to access the relevant data in the correct situations. Today, this is all handled autonomously by the AI, so there are zero decision trees or canned responses to maintain. Once APIs were exposed, Decagon's team took on the engineering lift to make the integration production-ready. Decagon's admin portal also provides a self-serve way to train the AI responses in real-time in the case of product updates or releases.

Throughout this process, Decagon's offering included world-class engineering support and responsiveness. This enabled the Bilt team to move quickly and confidently, especially while navigating the complexity of their tickets.
"Decagon’s technology has allowed us to save 1000s of support ticket and interactions month over month. Combining our internal data with their AI solution, customers can now received responses and answers to any enquiry in seconds. We’ve seen a dramatic rise is CSAT, NPS and our team can now focus on specific escalations and the broader user experience."
Thatcher Foster
VP, Client Solutions at Bilt
The Result
Decagon's new solution for Bilt has led to major improvements in all major metrics: deflection/resolution date, customer satisfaction, and time saved. The team is able to operate more efficiently, and with more insight into leveraging the support experience as a growth driver, and not just a cost center.

The hands-on, white-glove nature of Decagon's offering enabled the team to unlock powerful new ways to resolve tickets that previously required human intervention, with minimal engineering support from the Bilt team. A fully generative solution is the future for customer support teams, and Decagon is leading the way.
© 2024 Decagon. All rights reserved.