The spec is that they are supposed to cover 85% of the image area at infinity. That's definitely on the generous side, maybe more than most RF cameras -- but, I think, defensible considering that they have to cover a lot of different brands and types of lenses (which can vary surprisingly considerably from their marked focal lengths; Leica, for example, can be pretty confident in knowing how closely they mark their own lenses, and design their framelines accordingly, but Epson has to contend with Leica, Voigtlander, Carl Zeiss, Canon, etc., etc., all of which may have had slightly different standards.)
Do you think yours cover substantially less than 85% at infinity? I'm not sure exactly how I'd go about testing this, since infinity is such an inconveniently far-away distance: maybe find a subject with lots of recognizable details, such as maybe a cityscape or a large building, and compose (on a tripod) so that some easily distinguishable features line up exactly with the left and right side framelines; then, with the finished image open in your image editor, measure the distance between the two target objects and compare it to the total frame width.
It might be easier to get action out of Epson (if that's your goal) if you've got sample pictures to show them that the framelines are including significantly less than 85% of the final image. I don't think the frameline coverage is easily adjustable, since the frames are cut into a metal mask, but it would give you a talking point...