Where do the AI-Estimated data come from? How does it work?
Figures AI uses an open source machine learning model (XGBoost) to predict compensation for a specific combination of job / location / level. This predictive model has been trained on over the 90,000 real-time data points (within 88 countries) currently making up the Figures benchmark, considering the different roles, levels, and countries & locations. The model is trained over all percentiles (from 10th to 90th) in order to provide you with the whole range of compensation, for each role and level, as is provided by Figures’ traditional benchmark.
Why do I sometimes see AI-Estimated data and sometimes not?
Figures AI is displayed when we don’t have sufficient data to display the Fair, Strong, or Excellent data quality indicators. The AI-Estimated data aim to help you understand the compensation of a given job / location / level combination by providing a reliable estimate.
When there are sufficient real-time data, you will see the data quality indicators displaying Fair, Strong, or Excellent, alongside the number of data points and companies contributing those data points.