Introduction
Providing personalised and effective client service is crucial for the success of a service organisation. To accomplish this, it is essential to evenly distribute workloads among client advisors and make the most of each advisor's specialised expertise. This was the challenge we encountered while working with our client who managed a large and diverse portfolio of clients from numerous sectors. We therefore developed an allocation model to optimise the process of assigning clients to client advisors.
The challenge
Over time, client advisors found themselves managing portfolios that had become unevenly distributed. Some advisors were overburdened, while others had surplus capacity. Additionally, some advisors possessed extensive knowledge in specific areas, yet the clients requiring this expertise were distributed among advisors with lesser proficiency.
The organisation had to ensure that advisors with specialised expertise were matched with the appropriate clients without overloading any individual advisor. During the process of reassigning clients to advisors, it was crucial to minimise disruptions to the existing relationships.
The complexity of this task was further highlighted by the number of possible ways to allocate clients to advisors. Even with just 20 clients and 3 advisors, the number of possible combinations exceeds 133 million. Since the number of clients and advisors in this case was a multitude, the combinations grew exponentially, making it nearly impossible to solve through manual allocation or traditional methods.
Our advanced algorithmic solution
To address these challenges, we turned to a sophisticated technique called a genetic algorithm. This approach allowed us to search for an optimal solution that respected the constraints while maximising operational efficiency.
Genetic algorithms are inspired by the process of natural selection. They simulate different scenarios, evaluate their effectiveness, and gradually improve over time by retaining the best solutions. This method allowed us to explore a wide range of potential client-advisor allocations and identify the most balanced and effective configuration.
The solution was designed to prioritise parameters such as:
-
an even workload distribution
-
retention of sector expertise
-
language restrictions
While taking into account these constraints, the organisation’s longstanding client-advisor relationship had to be respected. It was therefore important that clients remained with their existing advisors wherever possible to ensure continuity of service. Our solution was designed to keep portfolio shifts to a minimum while still improving the overall distribution of workload and expertise.
As part of our solution, we also developed a user-friendly web application that allows our client to run simulations independently. This application enables the organisation to rerun the optimisation process using its parameters and priority weights for various constraints.
For instance, they can adjust how much importance they place on factors like workload balance, sector expertise, or client relationship continuity. This flexibility allows them to adapt the model to evolving business needs and priorities, ensuring that they can continually refine their client allocation strategy over time.
Conclusion
This case highlights the power of advanced evolutionary algorithms in addressing complex business challenges. By using these types of algorithms, we were able to solve an intricate problem in client portfolio management, improving both internal operations and client outcomes.
Our AI and machine learning expertise enabled us to deliver a solution tailored to the organisation's unique needs, balancing numerous constraints while driving significant efficiency gains.
