Using Voximplant Kit’s no-code drag and drop visual builder, two scenarios for calls to couriers were set up.
Scenario # 1: Calling a courier who has not used the app in the last 14 days and has a debt. When the voicebot calls, it introduces itself as a Glovo voice assistant, addresses the courier by name and says that the company has noticed a lack of activity on his or her account. The robot then asks if the courier plans to continue using Glovo. If the answer is affirmative, the robot reminds the courier that for using Glovo he/her needs to book delivery slots in the app. If not, the robot says that Glovo will always be pleased to get him/her back. Additionally, regardless of the answer, a request is made to refund the cash received from customers.
An additional personality validation mechanism triggers in case a courier claims that they are not the person the company is looking for. If triggered, the scenario plays first name and last name of the respective courier and repeats the question. If a courier continues to insist, the script marks this call for later analysis and finishes the conversation.
Scenario # 2: Calling a courier who has a debt and whose account has been deactivated due to a violation of the terms of service. During the call, the voicebot introduces itself, greets the courier by name, notifies that the account is blocked, and specifies the amount of outstanding funds that must be refunded by a certain deadline.
Speech synthesis and recognition technologies are used in the communication between a voicebot and a person. Immediately after the call ends, couriers receive additional information as well as a payment link via SMS and email. The voicebot uses data from Glovo’s CRM to dynamically personalize voice messages as well as texts for SMS and email (e.g. courier's name, debt sum, currency).
The project currently operates in seven languages: Spanish, Portuguese, Russian, Romanian, Ukrainian, Georgian, and Italian.