A couple ideas using awk or sed:

$ echo FT0215202301.xml | awk '{print substr($1,7,4) substr($1,3,4)}'
20230215

$ echo FT0215202301.xml | sed -E 's/^..(....)(....).*/21/'
20230215

Same solutions but using a here string to eliminate the subshell:

$ awk '{print substr($1,7,4) substr($1,3,4)}' <<< 'FT0215202301.xml'
20230215

$ sed -E 's/^..(....)(....).*/21/' <<< 'FT0215202301.xml'
20230215

If the filename is stored in a variable then you can use bash's substring capability (${var:start:length}) and eliminate the need for other binaries (eg, awk, sed), eg:

$ fname=FT0215202301.xml
$ echo "${fname:6:4}${fname:2:4}"
20230215

NOTE: bash string indexing starts at position 0

While the above solutions generate OP’s expected output of 20230215, OP’s cut -b 7-10,5-6,3-4 would seem to indicate OP may want 20231502 as the expected output. If this is the case the modified commands would be:

$ awk '{print substr($1,7,4) substr($1,5,2) substr($1,3,2)}' <<< 'FT0215202301.xml'
20231502

$ sed -E 's/^..(..)(..)(....).*/321/' <<< 'FT0215202301.xml'
20231502

$ fname=FT0215202301.xml
$ echo "${fname:6:4}${fname:4:2}${fname:2:2}"
20231502

I’d use Bash’s substring expansion:

f=FT0215202301.xml; echo "${f:6:4}${f:2:2}${f:4:2}"

20230215