Top Menu

The power of predictive analytics in workforce allocation

The Customer Experience & Management Transformation (CEMT) division at Telekom Malaysia (TM) is responsible for implementing solutions to continuously improve TM’s customer experience.  This includes the advancement of operation support systems, training of field officers and improvement of overall morale of field operators and front liners.

With thousands of customer orders to be evaluated every hour, it was imperative to match the company’s manpower and resources to attend to all customers’ calls at the most optimum time and cost considerations. Using multi-sourced data and Intelligent Resource Optimization Algorithms (IROA) to manage requests and the dispatching system, TM has been able to improve its service delivery by improving matches based on priority, constraints and resources.

Efficient planning and scheduling of workforce is important in telecommunications sector so that all work orders and customer demands per service-level-agreement (SLA) can be fulfilled in the shortest possible time. On-going huge demand for broadband services has led to an increase in TM’s investments in new equipment and resources. The increasing demand for installation of new infrastructure or maintenance will lead to an overwhelmingly large number of possible courses of action and ways to allocate the resources. One of the main objectives of TM is to ensure the service given to their customer or subscriber complies with the SLA. When any complaint or trouble ticket cannot be resolved via calls to the service centre, a maintenance team will be dispatched to perform an on-site evaluation and restore the service to meet the SLA target.


Thousands of service orders which are accompanied by millions of computed permutations need to be evaluated on an hourly basis using large amounts of data from various sources to meet the available resources capacity. Currently, this resource matching process is being done heuristically, which does not produce optimum solutions and this leads to resource wastage, for example:

  • Non-optimized waiting and idle time between jobs
  • Longer than needed travelling time between jobs
  • Inefficient selection of daily job dockets
  • Reliance on subjective decisions by supervisors

TM’s Workforce Management Allocation Optimization (WMAO) uses multi-sourced data and Intelligent Resource Optimization Algorithms (IROA) to manage the request and dispatching system in order to optimize trouble ticket flow based on priority, variable constraints and resources capacity. Workforce allocation is optimized based on criteria such as customer request aging, urgency of request, customer’s request for an appointment, travelling time and distance between jobs, skillsets required and working hours.

IROA runs every hour, every working day, nationwide, to assign jobs and teams based on location (building geographical coordinates) by considering the temporal and procedural constraints such as docket in hand and skillset required.  This is based on a scoring weightage in determining high priority jobs.


Business Benefits:
The IROA is an engine-based genetic algorithm created to overcome the previous system of manually managing large numbers of repair teams and cases. With the IROA, TM saw a reduction of operational cost by minimizing waiting time between jobs. There was also a decrease in operational cost by reducing travelling time between jobs, better sorting of daily job dockets, and elimination of subjective decisions from supervisors. SLAs were fulfilled and assignment of work was based on the most suitable skillset.

The Future:
From the current implementation of WMAO, it can be seen that the IROA utilises large data sets of relatively static (except aging) data. To further improve TM’s performance, live measurements can be utilised which include:

  • Customer churn risk
  • Live measurement of Network QoE level (attenuation, SNR)
  • Various external and relevant data sources (for example, weather and real-time traffic conditions)

, , , , , ,

Powered by WordPress. Designed by WooThemes