Rated as an “Excellent” on Trustpilot , BTC Direct is one of the most popular cryptocurrency companies in Europe. The customer service team receives on average 4,000 messages per month, via its crypto trading app, BLOX. However, when there’s a rapid change in the market, which is common in the crypto world, those numbers almost double and reach +7,000 messages per month.
The cryptocurrency market is volatile and there’s a direct correlation in the volume of incoming messages for the BLOX team. But it’s impossible for any company to scale its customer service team to double in one month and to half in the next. This uncertainty means that there’s either too much or too little work for the customer service agents.
Unlike traditional financial traders, crypto traders don’t adhere to the 9:30-16:00 stock exchange hours. When users want to buy and sell crypto, they want to do it immediately, even if it’s at 2:00 in the morning. If they have a question or there’s an issue, they also don’t have 15 min. to wait in a queue for a human to respond. They need the answer then, otherwise, it’s too late. Put simply, a delayed response to a user’s question results in lost sales for BLOX.
The other challenge that the BLOX team raised, is that the cryptocurrency is growing at an exponential rate and they need to scale with the market, without scaling their operational costs. In 2021, the number of people owning cryptocurrency increase by 178%! Crypto.com predicts that by the end of 2022, more than 1 billion people will own some form of cryptocurrency.
After reviewing the incoming user messages, we quickly agreed with the BLOX team that a chatbot is a perfect way for them to manage:
We took these insights and began building BiBi.
Following our chatbot scoping best practices, we focused on automating the high-volume, repetitive questions that were easy to answer. After reviewing the incoming user messages, we agreed on 20 intents that cover ~80% of all incoming questions.
Since BiBi’s scope is limited, we focused on making sure that she can answer her 20 questions almost perfectly. As a chatbot, she can handle multiple conversations simultaneously, it doesn’t matter if there’s one user asking a question, or 25, they all get their answers instantly. This means that we solve the users’ needs for instant responses regardless of volume, but we also we reduce the volume of incoming questions for the human colleagues so that they have the time and space to focus on more complex cases.
BiBi doesn’t work alone; she has a team of colleagues whom she supports. To ensure that the transition between her and her colleagues is seamless, we built specialized flows for opening and closed hours. Building flows when the customer support office is open allows BiBi to manage user expectations about how long it will take for a human colleague to take over the conversation. Having a specific flow for when the office is closed, means that BiBi can collect all the needed information from the user so that the human colleague has all the information that they need to respond to the user once the office is open again.
BiBi hands over the conversation to a human colleague during opening hours
BiBi collects user information when the customer service office is closed
We also noticed that the most time-consuming questions were usually related to issues with the single euro payments area (SEPA). SEPA is a financial system of standards and tools that facilitates easy cross-border payments. 45 countries are part of SEPA, and to buy and sell crypto on BLOX, you need to be a citizen of one of these SEPA countries.
Using entities enables BiBi to recognize when a user is or isn’t from a SEPA country
On average BiBi handles ~2,000 conversations end-to-end per month. She’s able to close 60% of the conversations in less than 1.5 min. and 80% in less than 2.5 min.
As the cryptocurrency market continues to grow, BiBi’s team will continue to grow her knowledge so that she can automate incoming questions from users.
CONVERSATIONS PER MONTH
SAVED PER MONTH
TO RESOLVE 80% OF CONVERSATIONS