As you probably know WOTC is introducing a new edition of D&D called D&D Next which is designed to accommodate a design philosophy called Bounded Accuracy (BA). I'll let you google BA.
BA should interest Paizo too, as a module developer and publisher, because there are three implications:
1. BA ensures that the "sweet spot" of gaming is spread across the entire career of character, not just within a narrow band of levels. This feature helps to keep players interested in playing Pathfinder(, or any RPG.)
2. BA increases the feasibility of playing modules outside the appropriate level range. This can potentially mean more revenue because the market of each module is widened.
3. BA eases the development of modules because the effect of inflation in Difficulty Checks (DC) due to gaining levels is depressed. This can potentially mean reduced cost of module development and more robust module designs.
BA can be implemented by making changes to the design of the game, which for Pathfinder is D&D 3.5. If that is unappealing to you, then there is a much easier way that also preserves the game design. Simply use a die that is larger than d20 when making roll checks, like d100 or 2d20. That's it. Get your mathematicians to investigate this implementation of BA.
For example, a requirement of BA is to increase the chance that low level monsters can hit high Armor Class (AC). As you know, AC is a type of DC. Suppose that a goblin with Attack Bonus of 0 is trying to hit AC 19. He would need to roll 19 or 20 on a d20 to succeed, which means his success rate is 10%. Rolling a d100, the goblin would need to roll 59 to 100 to hit the same AC which is now 50+9 (the 50 is based on the mid point of d100). This means that his success rate is now 41% > 10%.
I can make similar examples for Saving Throw and Skill Checks. Check out this implementation of BA. The optimum die to use seem to be 2d20.