Honestly (and I'm probably going to get strong disagreement on this), I think one of the biggest things holding back even the best game building software (like NWN) is too much reliance on the mindset and assumed skills of software design and coding.
A couple examples (may not apply to all systems, but is typical):
-Changes tend to be automatically applied, rather than relying on a save or apply feature. This assumes a higher level of competence than should be assumed, and is different from how most software for typical consumers operates.
-Insufficiently accessible UI. They should have lots of drag and drop functionality, better organized menus, etc.
What is really needed is an end user accessibility consultant with limited (though not zero) software design skills. They need someone to periodically look at it and give meaningful feedback on what is not as straightforward as it should be.
The systems could also be designed to do more out of the box, without any scripting or involved scripting-like design. While AI might help with it, it shouldn't be necessary.
For instance if you are setting up dialog options and other results of interactions with an NPC, it should pop up a visual flow chart where you can click to make new nodes, drag lines between them, and have other simple buttons you click to apply various functions (like giving or removing items, XP, quest stages, etc).
If you are setting up a trap, there should be a bunch of presets with easy to apply options. So you might grab and plop down a Proximity or Activation trap. When you plop down a Proximity trap, a sidebar (or floating box connected to the trap) could have a slider that lets you change the size of the triggering region and a few buttons (circle, rectangle, line) that you click to choose. The sidebar or box would also have buttons for damage type, and for amount of damage, or special effects like teleporting to a location, applying a condition, etc. (The special effects may need to open a separate box when clicked to show buttons with the various options, but the rest should fit in the sidebar or floating box). You always use visual buttons, and either include text with them, or have text appear when you hover.
And the whole thing needs to be that easy. And it really can be. I think people are probably going to be skeptical because they haven't seen it done right before. (I'm even willing to put my money (or in this case, time) where my mouth is and do the job myself of being such a consultant (free to start and prove it's possible, though I wouldn't work free forever.))
But it is always going to struggle as long as the people making it are primarily making tooks for their own use. They need to be thinking "How can I make this tool so accessible and good at what it does that I put myself out of a job?" (which it wouldn't actually do) and be willing to put in the effort to actually listen to the consultants and do things in ways that are inefficient for themselves.
Obviously, most toolkits that are attached to a specific game aren't going to want to spend enough resources to do that. (Maybe making a paid tool would be a better option.) But even dedicated game design software doesn't really do what I'm talking about. Various software has bits and pieces of what is needed, but it really needs to all be put together; and it's probably going to work best if it is built to go with an actual game and popular RPG system (assuming an RPG design toolkit here).
So that's what the issue is. I've played around with these enough to see how they could be made much faster and easier so people could just basically think of the story they want to create and quickly make it happen, rather than having to spend enormous amounts of time messing with the software to figure out how to do it and painstakingly implement it.