Hampshire, UK-based Neural Technologies has unveiled its new telecommunications customer churn predictor, DA Churn Manager which uses neural technology to analyze historical customer data and predict if and when, a customer is likely to churn. It can also predict the propensity of certain customers to churn, and predict the likely effects of forthcoming promotional campaigns. DA Churn Manager is one of Neural Technologies’ three flagship products. The other two are Decider, for consumer credit scoring and risk analysis, and Minotaur, launched last year (CI No 2,957), which looks for mobile phone fraud by building a profile of a user’s typical phone usage, and then using neural techniques to spot abnormal usage patterns. DA Churn Manager uses a computing architecture developed by the company called AMAN, Advanced Modular Adaptive Network, that enables Neural Technologies to combine the right type of neural network model or models with, in some cases traditional computing techniques, in a modular fashion, to produce the best tools for a given application. The DA Churn Manager development environment is a stand-alone application which runs typically on a Windows NT workstation and enables churn models to be built and tested before being deployed. The models are then deployed using the AMAN server, which runs on NT, Unix or OpenVMS. The system is currently being given trials by a number of companies in the UK, but no names were given. Peter Baxendale, commercial director for Neural Technologies, says the system has been consistently forecasting ‘churners’ with an accuracy rate of more than 80%. The company is now talking to service providers and network providers in the US, and says feedback is very positive. Founded 10 years ago, Neural Technologies specializes in building vertical market applications using neural technology. Last year, the company opened an office in Singapore to go after the South East Asian market (CI No 2,957), and says it is now seeing healthy trade from that region, as well as Australia and South Africa.
