As always, folks have different preferences as to methods, based upon personal assessments of risk-benefit analysis. This got me to thinking about two variations that seem reasonable and may be worth consideration.
One is a simple change to the impact of a 2NT (poor trumps) cue. One might agree that any cuebidding sequence that includes a 2NT call can end at 3NT. Thus, for instance, consider 1♠-P-2♣-P-2♥-P-2♠. If Opener cues anything but 2NT, the contract would be forced to 4♠ (if game only). However, under the variation, a 2NT call could yield a final contract of 3NT.
A second variation, along similar lines, might be to invert the meanings of a Picture Jump to 3NT and a Picture Splinter in partner's second suit. Thus, after 1♠-P-2♣-P-2♥-P-2♠, Opener might jump to 3NT to show a stiff club, good trumps, great hearts, and no diamond control, a call that Responder might actually want to pass. That would make a 4♣ call by Opener instead show good trumps, no heart control, a stiff diamond, and HHx in clubs, a holding where 3NT is very unlikely to be right.
A third possibility I have considered is also to invert the meanings of 3NT and 2NT, where 2NT is "serious" and bypassing 2NT "non-serious," or perhaps frivolous and non-frivolous, with 3NT being the "poor trumps" cue. The upside is for those who very highly value non-disclosure when a quantitative bash analysis suggests that slam is remote. The downside is that it negatively impacts the nuances available for delayed picture jumping. But, that itself might be worked out sufficiently.