As of today, our Excel export feature uses the type inference mechanism of pandas. This explains why the ID column was converted to a numerical one in the file export.
Thanks for pointing it out, we will see how we can improve this in future releases.