Predictive Dialer Integration

Advancing call center technology

Technologies:

Java, Spring, Hibernate

About the project
Making of

In the processing of any call-center the operator has two main factors that waste his time:

  • Waiting for a call to answer: the time before the client pick up the phone and start talking.
  • The time required for a new call in case of previous one was rejected. Relevant especially for debt collection calls.

The developed solution has minimized the response time and reduced downtime.

A predictive dialer is a call processing system based on complex probabilistic and statistical modeling. The solution predicts when the next call is likely to get connected, and when the next operator is likely to become free to take a new call. This is a ceaseless process, so there is zero wait time for each call. A predictive dialer starts dialing numbers a few seconds before an agent has finished a call so that the operator can immediately be on the next call, without any downtime. In this way operators can spend their maximum time speaking to customers rather than trying to call to the next number on the list, so more calls can be made in the same time resulting in an overall boost in the call center productivity.

Depending from the number of operators the predictive dialer distributes calls in such a way that all employees are loaded equally throughout the day. At the same time the solution has a possibility to take into account the number of factors (one can speaks faster and his average call processing speed is higher) collected in the past, and those foreseen to happen after the call is made, to maximize the productive time of operators.

Knowing about the duration of the call, the number of answered and unanswered calls, the number of incoming and outgoing calls for the selected period of time, you can find out how efficiently this or that direction of calls is, where and when the main time losses occur.

The advantage of custom solution is that it shows a performance gain for the group of 5 operators already (industrial solution benefits only at the level of 25-50 people ). And at the same time, custom solution can be adjusted to any software implementation of IP telephony and integrated into any CRM (industrial solutions come bundled with both telephony and often with CRM). Besides the development costs less.

Results

We implemented this custom predictive solution for SMSFINANCE — a fintech-company from Akademgorodok, Novosibirsk. The company provides online financial services for individuals since 2012. Services are available in all regions of the Russian Federation. Today SMSFINANCE is a Russian top-5 online microloan provider. As a result the client has dramatically increased the call center productivity and improved campaign bottom-line. It was made possible by predictive dialing, skipping busy, no-answer and disconnected lines, and most importantly not involving the operator on a call until a live person answers.

  • A call center is able to process 35% more requests at the same time.
  • The solution shows a great performance even at "low" time (e.g. at nights with 10-15 operators online). At the same time enterprise solutions shows effect only with 25 operators involved.
  • Debt collection efficiency has increased threefold. On the first day of the work, the predictive dialer called the debtors scheduled for the day by lunchtime.

Meet other comprehensive fintech projects made by our mobile development team:

Project team

Andrey Tarasenko

Project Manager

Olga Ippolitova

Java Developer

Danil Kadetov

Java Architect

Dmitry Shesterikov

Lead Java Developer

Learn more