🚀 go-pugleaf

If you record in edit view, with levels set so that when they talk at the most loud it doesn't overload the recording, you should be able to use compression ('dynamics processing' or 'hard limiter', see manual for details) to even out things afterwards, even though to begin with some of the speech was pretty faint. That's assuming you have low noise mics and low noise preamplification - of course there is a risk that if they talk very quietly the noise of the mics and preamp may be greater than the noise of the voice if the equipment is poor.

You could also try a piece of free software on the recorded files - "The Levelator" which google would find.

However, if the mic technique is really poor, that means they could be way back from the mic, and you get as much coughing, traffic noise, and aircraft noise as you get from the mic. If that's the case, then the louder you make the speech, the louder you also make the background noise, and there's no automatic process that will really help you.

What you may have to do is to initiate some training of the speakers. Or have someone in the front row who will quietly come forward and adjust the position of the mic if required during the reading.