Real-time / proactive optimization.

This module of Incelligent’s software framework delivers innovative knowledge-based/proactive recommendations, on how to handle the addressed/anticipated situations. The goal is to optimize key areas of the Telco business, as indicatively customer experience and the operational expenditure. 

The module is based on three fundamental functions:

1. Modelling of:

  1. the business goals (objective), and,
  2. the available options / actions (repertoire of policies)

of the client, typically a network operator.

2. Leveraging big data and predictions that represent past and future conditions.

3. Evaluating the impact of each option in the repertoire of policies (actions) and selecting the optimal.

The output of the module is a stream of real-time recommendations with the actions that optimize the objective. The recommendations can be automatically enforced on the network, or be provided as advices to an operator.   

The product manages multi-vendor infrastructures, which may also be heterogeneous, by including 3GPP, Wi-Fi and other network technologies.

The repertoire of actions can address network management aspects or customer facing processes.

Indicatively, in the network management case the output typically affects the network topology, the spectrum allocation, the network behavior, and the traffic management. Network topology changes can involve the activation/deactivation of small cells and the alteration of the power levels transmitted. Network behavior changes may involve the selection / activation / deactivation of SON mechanisms and the recommendation regarding respective parameter configurations.

In customer facing operations the output addresses challenges in the area of customer care, retention, and sales. Indicatively it may affect parameters of promotions, human resource allocation in time and space (e.g. fault management) and proactive communications and compensations to customers. 

The core function

In order to produce its recommendations, incelliOpt continuously learns the performance of the application of certain actions in the network/subscriber base. Thus, it builds a historical base with the actions taken in a specific contextual situation in the past and the performance they resulted to. This knowledge is stored, and can be looked up (pattern identifications) and selected to be enforced, whenever a similar contextual situation (e.g. in terms of traffic, mobility, time, weather, etc.) is predicted by incelliAna.