Securitas improves sales force planning with an intelligent rule system

The application of intelligent rules to optimize processes and artificial intelligence to improve decision-making is a growing trend. The world's largest companies are already employing algorithms and intelligent models to leverage large volumes of data and support their decisions.

Securitas Direct, one of the best-known security companies, opted to apply this type of rule system to enable scheduling the visits of their sales representatives according to key business indicators. The aim is none other than optimizing their representatives’ travel routes when visiting real or potential customers. The initiative was prompted by the company's innovation department, which counted on the support of Sngular's Data & Artificial Intelligence team from the start, at a technical and functional level.

Nieves Velásquez, engineer and backend developer, and Jorge Prudencio, cognitive solutions architect at Sngular, kicked-off this pilot project, analyzing Securitas' starting premise and needs. In addition, Nerea Luis participated as a data analyst, Mariano José Zardoya as a project lead, and Manuel Bago as a business manager. After defining the general concept, the team made their first proposal, which was improved based on the feedback received from the parties involved. Then, the development of the solution began.

From that moment on, the team scheduled biweekly meetings to present demos of the new developments and plan further improvements, maintaining constant communication to ease the progress of the project. Three months after the start, Sngular’s team presented the first version of the service and Securitas Direct's AWS production environment was deployed. Then, the first production tests began.

Although this is a pilot project, the first real tests will soon be launched within the company to gather feedback and define a wider production release.

Smart rules to enhance productivity

The solution’s aim is to support planning departments by establishing intelligent and automatic workload distribution rules. The system makes it possible to optimize and balance sales reps’ visits based on factors such as the distance from the visting point, team schedules and workload, among others.

How does it work? The service consists of three steps:

• The KPI information of each branch is loaded daily in an auto-generated file, together with a list of backup branches that will receive the reassigned visits.

• Then, the service receives the zip code of each reassigned visit with a unique identifier.

The system searches for the KPI information associated with the zip code and, by applying a series of intelligent rules, checks whether there is a backup branch that can handle the visit.

If the system finds a suitable branch, the visit is assigned (and its related information is sent); otherwise, if it’s not possible to reassign the visit, the system returns the information related to the branch the visit was initially assigned to.

• Lastly, another service generates business metrics for further analysis.

In the second phase of the project, Sngular’s team helped Securitas Direct improve the system for obtaining alternative branches by using maps (GIS technologies), taking into account both the sales reps’ travel time and the customers’ distance. This reduces time on the road and optimizes visits.

Sngular's collaboration with Securitas Direct does not end here: in addition to this project, we are working with them in other areas such as QA.

How to translate business logic into intelligent rules

For Sngular’s team, the biggest challenge of the project was to create an intelligent system to apply the daily assignment rules provided by Securitas Direct. Once the system was designed, they needed to find a tool that would permit to implement the logic in a scalable way.

Setting up the infrastructure on Amazon Web Services - the project’s cloud provider - has enabled the team to widen its knowledge, since they made possible to deploy all the resources configured and synchronized in the different work environments by importing a file in CloudFormation. All this permitted the team to automatically create and configure such resources and export them if necessary.

Node.js has been used to develop the assignment rules’ logic, and this logic has been built on top of and a library called json-rules-engine.

In addition, GIS map visualization technologies were used to visually represent large volumes of data coming from Securitas Direct repositories. Data Visualization technologies are also growing and becoming key in recent times; an example of it are the visual maps used to display the spread of COVID-19 we are all familiar with.

In the words of the people involved, working on this project with Securitas Direct has been "undoubtedly challenging and rewarding, due to the complexity of the rules and of the infrastructure we needed to create".