My intuition is that you could use these techniques to prove that if things are ontologically foundational but only coherentism is epistemically justified, then in the limit as you explore everything your coherent knowledge approaches the foundational reality (subject to a variety of assumptions which are unlikely to be true or at least provable in practice).

In particular, you'd probably need to do something like parameterize statements p with their complexity and compute weighted error between p-as-justified-by-coherent-belief and ground-truth-foundational-p over all statements p; complexity would need to be parameterized parameterized such that total error is finite (i.e. weight drops, in the limit as statements grow arbitrarily large, enough faster than the number of possibilities increases).

I'm not sure what the point would be, however.