I'm familiar with Fermat's Little Theorem and Euler's Totient, but I'm wondering whether the fact that the only shared factor of $(a,N)=1$ has something to do with the fact that, given the prior constraints there exists at least one $x$ (with $x$ being one of elements of the least non-zero positive remainder class of $N$) where $a^x \equiv 1 \pmod{N}$?
It doesn't appear to be dependent upon the modulus being prime since using a modulus of 10 the following is true:
$9^1\pod{10}=9$
$9^2\pod{10}=1$
$9^3\pod{10}=9$
$9^4\pod{10}=1$
$9^5\pod{10}=9$
$9^6\pod{10}=1$
$9^7\pod{10}=9$
$9^8\pod{10}=1$
$9^9\pod{10}=9$
This is also really naive but another observation is that given the previous assumptions the initial base will always be less than $N$ and $a^1 \pmod{N}$ will always return $a$, so given $\gcd(a,N)=1$, $a$ will always produce at least one remainder other than $1$.
I know there's some intuition that I'm missing?
Consider all powers $a^r\bmod N$ for $r=1,2,3,\ldots$ There are an infinite number of them, but they can only take values $0\le a^r\bmod N\le N-1$. So sooner or later we must encounter a repeated value; say $a^s\equiv a^t\bmod N$, for some $s<t$.
Now, since $\gcd(a,N)=1$, we know that $\gcd(a^s,N)=1$. So we can divide by $a^s$, and we get $1\equiv a^{t-s}\bmod N$. Which is what we wanted.