I am currently working on a time series task related to predicting the AI demand in the market. I have already obtained the monthly percentage of AI jobs relative to total jobs from 2018 to 2023. I also want to include some economic data as exogenous variables to help improve the forecast model. I have a question: Is it necessary for my exogenous data to include a future period (such as 2018-2024)?

I have use SARIMA and prophet to do it, but I want to try if the model performs better with exog data.