This, among other things, is a response to Frege's claim that concepts must be fully determined and have a value for every object, and a response to his former self, who thought that "that if anyone utters a sentence and means or understands it he is operating a calculus according to definite rules" (PI §81). But there are two ways to understand what Wittgenstein is doing in §88.
a) He is arguing that some things, concepts and orders included, have fuzzy inexact boundaries instead of exact ones.
b) He is urging us to examine our notion of exactness--what exactness comes to in different cases, and what reasons and occasions we have for using notions of exactness.
The first kind of reading is the more popular--it is orthodox. Anyone who "knows" something about Wittgenstein "knows" that he said our concepts have fuzzy, inexact, boundaries. This orthodox reading brings with it a host of philosophical temptations: careless relativism for one, but more generally a kind of indifference and slackness: 'we don't have to be exact, we can just call it the way we see it, and that'll be good enough.'
But it is not true to Wittgenstein. He writes a little bit later in the same section (§88): "'Inexact' is really a reproach, and 'exact' is praise." And Wittgenstein can hardly be understood to be reproaching our concepts or orders. Similarly, questioning the idea that we operate by definite rules (§81) is not arguing that we operate by indefinite rules.
The first reading takes for granted something that Wittgenstein is putting under question. Wittgenstein is not, and he is not taking himself to be, in a position to say that our concepts have no exact boundaries, or what an application of a word that is everywhere [absolutely] bounded by rules would be like, (§84). And that’s because he does not take himself to be clear about the notion of our concept having exact boundaries, or even what it means to proceed by a rule (§82). That is, in order to be able to say "our concepts have inexact instead of exact boundaries," he would have to already know what having exact boundaries mean.
That is exactly what his investigation is about. That is, it is not something he takes himself to be in a position to assume. §88 is among other things about investigating the (grammar of the) word "exact." And what Wittgenstein wants to signal is not that we are mistaken about something (that we took our concepts to have one kind of boundaries, whereas they have another), but that we have a fantasy of something—a fantasy of exactness. And he is not saying that exactness is itself a fantasy. Rather, there are fantastic notions of exactness, and realistic notions. To get at the realistic notions, he suggests, we need to look. We will tend to keep reaching for fantastic notions as long as we don't look. And this is exactly what Wittgenstein is doing. In §88 we see Wittgenstein in the process of looking. And as a consequence, he is not there in a position to deny anything--to deny exactness to our concepts.