AI-powered workforce management in a Brazilian health network

Aussteller: XIMES GmbH

„Diagnosticos da America“ (DASA) in São Paulo, Brazil, is the largest health network in Brazil with over 40,000 employees in walk-in clinics, labs, and 14 different hospitals. In the labs and call-centers, workforce management is difficult: Work demands varies over time, labor laws need to be considered, and multiple functions and qualifications must be differentiated for personnel planning and scheduling. Previous workforce management with Excel was limited in data capacity, forecasting personnel demand, and reviewing and validating historic data. Scaling the workforce according to the forecasted demand, accounting for adequate cover and legal issues, was challenging. Employees were unsatisfied with overtime and understaffing during rest breaks.

Therefore, a new software solution was developed and implemented to improve scheduling and employee satisfaction. The software TIS – Time Intelligence Solution by XIMES GmbH is a business intelligence software using AI methods tailored to handle time-based data. The partner PortiaLAC was trained and adapted the TIS solution suite to DASA’s requirements.  DASA planners can autonomously adjust the system to changing circumstances.

90% of the clinics and labs and 31 units of call-centers within DASA with about 2,500 employees were included in the forecasting and scheduling of the workforce. Three components were implemented:

TIS PROGNOS: This solution calculates a forecast of personnel demand based on actual historic data of calls and visitors to labs for ca. 800 planning areas. Calendrical data and influencing factors are included in seasonal regression models to predict daily and hourly personnel demand. Modeling includes validation of the model fit, and statistical relevant influencing factors such as lock-downs during the COVID-19 pandemic.

TIS DEMAND: Based on the forecast, a simulation approximates the minimal personnel needed to fulfill a service level in the call-center. A complex definition of service level can include e.g., a waiting time of less than 20 seconds for 80%, and less than 60 seconds for 90% of all calls. The results improved the previous service quality with only a small increase in personnel.

TIS Operating Hours Assistant (OPA): In this step, the output of TIS DEMAND is used to find and optimize adequate work shifts to cover the demand. Several legal and company-specific rules for scheduling can be applied. As an example, specific rules for taking rest breaks were not considered previously, leading to understaffing at certain times. With OPA, requirements for work shifts including rest breaks could be considered, decreasing the understaffing.

The new software solution greatly improved workforce management quality. Planners were supported with adequate tools, and employees benefited from less overtime and better work schedules. The system allows to calculate different scenarios and planners can select best fitting scenario based on a comparison of the results.


Name: Christian Steindl
Telefon: +43 1 535 79 20-0