An Outbound Algorithm for Dynamically Adjusting Call Frequency

Abstract

The invention discloses an outbound call algorithm for dynamically adjusting call frequency, comprising the following steps: extracting a plurality of historical call data samples in different time periods, and marking whether the data is connected; performing secondary processing on the marked data, Calculate the outbound call completion rate and call loss rate per unit time of connection data, and use it as the expected input value; build a three-layer neural network, input the historical call data samples and expected input values ​​into the neural network, and calculate the actual call value per minute; Compare the actual call value with the expected input value, reversely adjust and optimize the connection rate and call loss rate, and input the adjusted and optimized connection rate and call loss rate as the new expected input value into the neural network unit, and calculate the maximum rate per minute. Optimum call value; predict the optimal call value per minute for outbound callers, initiate test-type batch outbound calls based on the current number of online seats and current idle seats, and dynamically adjust the frequency of outbound calls according to different project models.

Description

technical field [0001] The invention belongs to the technical field of call centers, and in particular relates to an outbound call algorithm for dynamically adjusting call frequency. Background technique [0002] At present, for a large number of outbound business projects, there are two common methods. One is to use hardware equipment to make outbound calls: one agent corresponds to a dialer. Receive the agent; if the customer does not answer, then hang up and call the next call until the customer answers, and then transfer the call to the agent. In this method, the dialer locks an agent, and if the agent is free, then dials to help the agent call the customer. If the calling customer is always connected during this process, the agent is always in a waiting state, which is inefficient. [0003] In addition, in the prior art, it is also proposed to predict the number of outbound calls by using a prediction algorithm, and to allocate dials based on the number of outbound cal...

AI Technical Summary

Problems solved by technology

In this method, the dialer locks an agent. If the agent is free, it will dial to help the agent call the customer. In this process, if the calling customer is always connected, the agent is always in a waiting state, which is inefficient.

[0003] In addition, in the prior art, it is also proposed to predict the number of outbound calls by using a prediction algorithm, and to allocate dials based on the number of outbound calls. However, the success rate of outbound calls and the duration of calls are related to the active behavior of users, which cannot be accurately predicted, resulting in the prediction results There is a gap with the actual situation, so that it cannot be close to the actual situation for outbound dialing and scheduling

And there are already methods on the market for calculating the forecasted outbound call volume of call centers, but some of these methods predict inaccurate call volumes, some cannot be dynamically adjusted according to real-time call situations, and some algorithms are very complicated

Images