The following distribution of requests was observed
Posted: Sat Feb 22, 2025 6:15 am
August 2019, Dinero faced the urgent task of unloading operators, implementing a chatbot for initial customer requests, and automating the processing of typical questions.
LiveTex, which had successfully implemented several similar projects before, was chosen to develop this project. A great deal of help was provided by the friendly company Metrocredit , which already uses Neuromodule .
To launch the project, it was necessary to prepare a list of the most popular requests, answers to them, and also train operators to work in the chat so that they could promptly respond to customer requests if the chatbot could not answer the question.
After all the settings and anhui mobile number database testing in a neutral environment, it was time to “roll out the release” to the “combat” site.
Let's look at the key changes in the first month after implementation:
The widget was displayed 230,395 times.
Visitors clicked on the widget 21,281 times. Thus, the conversion was 9.24%, which is an excellent indicator.
Out of 21,281 requests, 4,993 were resolved using the chatbot. The conversion rate was 23.4% - every fifth request was processed without the participation of a real operator.
The remaining 16,288 requests were distributed to operators and, given the reduced workload, were processed much more efficiently than before - the number of missed and unprocessed chats decreased by 30%.
LiveTex, which had successfully implemented several similar projects before, was chosen to develop this project. A great deal of help was provided by the friendly company Metrocredit , which already uses Neuromodule .
To launch the project, it was necessary to prepare a list of the most popular requests, answers to them, and also train operators to work in the chat so that they could promptly respond to customer requests if the chatbot could not answer the question.
After all the settings and anhui mobile number database testing in a neutral environment, it was time to “roll out the release” to the “combat” site.
Let's look at the key changes in the first month after implementation:
The widget was displayed 230,395 times.
Visitors clicked on the widget 21,281 times. Thus, the conversion was 9.24%, which is an excellent indicator.
Out of 21,281 requests, 4,993 were resolved using the chatbot. The conversion rate was 23.4% - every fifth request was processed without the participation of a real operator.
The remaining 16,288 requests were distributed to operators and, given the reduced workload, were processed much more efficiently than before - the number of missed and unprocessed chats decreased by 30%.