The SQL Server team was thinking about AI built-in before AI was a “thing.” In SQL Server 2016 we introduced SQL Server 2016 R Services. This feature allowed developers to run an R program in a sidecar fashion with the SQL Server engine. When you installed this feature, we deployed a separate set of services on Windows to allow an R program to be integrated with data in SQL Server and execute code to run a machine learning model. Internally we called this feature SPEES, which stands for sp_execute_external_script. This is a system procedure you used to run your R code you installed locally on the VM or computer where SQL is installed. Our thinking was to bring the world of data science to SQL Server instead of running on client computers. SQL Server could be used to securely run the R code in a very scalable way. We even supported resource governor workgroups to control resource usage.