Counting, the way you are used to it, isn't real. I mean, it makes lots of sense in our current low-entropy universe, where distinct things that are pretty similar are around, and we have lots of brains to notice it.
But 1 cookie, 2 cookies, 3 cookies -- that isn't a fundamental thing about our universe. There is a bunch of stuff. When ridiculously highly structured, as a shorthand, you can assign different stuff to have a similar label, and then count how many things have the same label. And there isn't only one kind of thing you can count, but more than one kind of thing! (Even more structure). And when you say two things have the same count, you can move them around (more entropy generated) and put them in correspondence to each other.
Counting is a common pattern in our low-entropy current universe epoch. Making the abstraction of counting -- 0, 1, 2, 3 and so on -- mostly lies on top of the fact that there are piles of things to count, and we know how counting things work.
Our universe's current state is ridiculously highly ordered because of how recent the big bang was. Entropy hasn't had time to grow to turn everything to a smear -- so while that is true, there are going to be patterns, and those patterns are going to be pretty similar to each other, and those similar patterns are going to be countable.
Once you say "there are 1 thousand sheep", and have an idea what people mean by "sheep", you can encompass a ridiculous amount of information really quickly. Given 1000 sheep, you know you could split them into pairs, and take the left and right halves, and each half would have 500 sheep.
Expressing that without counting would involve understanding and knowing each of the 1000 sheep as distinct things, individually talking about the concept of pairing and left/right for each of them, and then understanding each of the piles of things (called "sheep"). A real pain.
Mathematics acts as a kind of compression. We label things as sheep (a category of kinds of stuff clumped in a particular way), say we have 1000 different stuffs that can be labelled as them. That is much, much more structured than "we have 20,000 kg of various proteins, fats, minerals, liquids and carbohydrates arranged in this specific way". (Note I used a number there, hard to get around).
If you accept that -- that mathematics is compression, or shorthand, that lets you talk about patterns of various kinds in much cheaper ways -- then the rest of mathematics falls out.
What is $i=\sqrt(-1)$? Why, it is yet another pattern. When you start with the counting numbers, you can then find the pattern of fractions. This pattern can be used to express things even more powerful than counting numbers.
From that you can find the pattern of the continuum -- the real numbers -- which again can be used to express even more powerful thoughts.
As it turns out, certain things can be expressed using polynomials; "x squared plus two x minus 3" for example. They are powerful tools that let you understand how things (in our highly ordered, low entropy universe) move, fall, and the like.
Those polynomials in turn are easier to work with if we invent a symbol we called "i", which when squared equals -1. It doesn't have to correspond to anything physical for it to be useful; in fact, in many situations its existence in a "solution" to a mathematical equation is strong evidence that there isn't a solution at all. But it merely existing makes finding the solution (or lack of it) easier; using the real numbers with "i" added (aka, the complex numbers) makes doing the mathematics (compression of understanding of reality) easier, and reality gets compressed better.
The mathematicians play the game of numbers carefully, and are pretty convincing that the addition of "i" doesn't break the game when played disconnected from counting.
So now we have these complex numbers. As it happens, you can find other parts of reality -- rotation, electrical potential, quantum mechanics and a whole pile of other things -- in which you can connect the complex numbers (including "i") to physical phenomena and patterns in ways that the complex numbers generate useful predictions of what happens next. They are good at compressing things. So they are useful mathematics (in the applied sense).
Are they "true"? Well, I'm starting from the position that counting isn't "true". They don't need to be "true" to express truth or be useful.
Because it is true that at least two yummy cookies are waiting for me at home, even if counting isn't really real.