Best way to get that vocal vowel is with a formant filter. You can take any waveform you want, run it through it through the formant filter and tweak it’s cutoff. Or the filter might have a dedicated control for choosing/shaping the vowel, depending on the synth you use.
If you don’t have access to a formant filter, cranking a low pass filter’s resonance, and carefully setting the cutoff, can also push out a vowel-esque character in the filter.
This track’s patch also has a noise layer. It sounds like it’s a separate patch, running through its own filter. It’s really bright on the first two notes, but gets shaped more like the pad on the 3rd note, and adds a breathiness that makes that 3rd note sound even more human.