You probably forget that Trump had set May 1st as his withdrawal date. Not that I think it would have happened then, but the draw-down for that would have been even more hasty (or incomplete).
A May 1st (is that the start of or end of?) withdrawal would've been under the condition that the Taliban not take any additional territory, which they had violated.
Let us assume that the Taliban did nothing to prevent the presumed Trump-led start of the process, which we shall assume goes ahead with measured retreat as the Taliban politely offer no action to jeapordise the agreement (not even rogue(/more-rogue!) elements, which have obviously been itching to rise up). The staged withdrawal you describe happens over months, all strangely but reassuringly peaceful, until it reaches the almost-totally-withdrawn status of just before our own timeline's surge-resurgance. We can see that the force needed for the surge-resurgance was there (or was built up) and would therefore have surge-resurged in even the perfect Trumpian world.
The only unbeliivable part of that scenario is that the underground mass of the (re-)occupying Taliban could be so restrained, having by practical necessity eschewed an official command hierearchy with ranks and organisational structure charts and all the guff that conventional armed forces do not need to so overwhelmingly keep obscured and fluid and versatile in case of counter-intelligence compromises.
Actually "gone completely quiet, causing no trouble at all" is probably a warning sign in itself. If there's suddenly not some Taliban cell causing trouble for some national army outpost somewhere, at times like , I'm convinced that a "trust but verify" approach would tread even
more carefully. With absolutely none of the remote Taliban equivalents of country-bumpkin yahoos causing ruckus, it suggests that someone (via the concealed command-structure) has put the fear of
GodAllah into them to keep the noise down. And for what reason are they all lying low? Perhaps because the same someone who is equally capable of prompting a coordinated uprising has plans (e.g., said coordinated uprising) for later.
It's far too early and far too fog-of-warred to really say what would have worked. I'm tempted to say 'nothing', including continuing whack-a-mole actions as before. The question "if not now, then when?" has been posited in multiple situations, this year, and strangely those who use it for opening up 'post' Covid (where we definitely know there's still reservoirs of illness itching to unleash themselves upon the population) seem also to be amongst the crowd inclined to the "not yet!" post-hoc rationalisation for Afghanistan (where we definitely know there's still reservoirs of insurgents itching to unleash themselvez upon the population). And, to be fair, there's those who have been working with exactly flipped supporting/hesitating attitudes to these two scenarios. And I'm sure plenty have changed their minds on either/both of these example points even if at least being equivolently ideological on such matters at some precious moment of time.
It's all above my paygrade, this is just me being a rando with a viewpoint. You seem to have the same limitations, though applied in a different direction. And perhaps a lot more self-certainty, I know.