Using the normalize effect will get your loudest sound to the highest level it can be on the CD. This is only the start. These day we are used to a compressed sound range on CD so the the quietest sound is un-naturally loud. This is very useful in a car where there is a high level of background noise, but we have become used to it in all situations. Which is why people who are used to CDs find LPs rather empty sounding, its kinda sad but it is as it is.

I wouldn't use the simplistic compressor supplied with audacity but use the add in dynamic compressor. The default settings work well for speech, try it on your speech track.

You also might consider giving the listeners some help in distinguishing the speech from the flute. One way to do this is to pan one track left and other right. If you had two music tracks you would consider those tracks left and right and leaving the speech in the middle.

The other way is to use the tonal range to split the speech and flute by EQing each track to fill a different part of the spectrum of frequencies. You might find the flute and voice already fill different frequencies and you just need to reduce the overflow into each others spectral territory.

Simon