Applications of super-mathematics to non-super mathematics Supergeometry and more broadly supermathematics has been around for few decades. Since its introduction by physicists, there has been an some mathematical interest in them.
Although interesting in its own, a classical mathematician (say someone who is only interested in smooth manifolds, complex geometry, representation theory, or number theory etc...) can say that these new super-objects are merely formal analogues that are only of interest to physicists, but not much to them.
One way to motivate such a classical mathematician to be interested in super-mathematics is to exploit some applications of the super world in the non-super world, and here comes my question:
What are the known examples of advances*/accomplishments in  non-super mathematics in which super-mathematics has been used in an essential way?
I am mostly interested in examples coming from Geometry (algebraic and differential), number theory (in the broad sense) and representation theory (e.g. Lie algebras or algebraic groups), but I am open to hear about other areas as well.
Here,  "advances" is meant in the broadest sense. It can mean a new perspective on an existing result/field, a new proof of a known fact or a totally new discovery etc...
 A: Etingof-Gelaki's classification of semisimple and cosemisimple triangular Hopf algebras (arxiv, journal), goes via Deligne's theorem about the existence of super fiber functors.
A: $\mathbb{Z}$-graded supersymmetry, where we have a $\mathbb{Z}$-graded ring $\{A_k\}_{k\in\mathbb{Z}}$ satisfying the law
$$ab=(-1)^{\deg(a)\deg(b)}ba$$
has recently been applied in the setting of homotopy theory in the on-going work of Hesselholt–Pstrągowski, started by Dirac geometry I: Commutative algebra.
From what I understand (very little), the main example and motivation for it is that the $\pi_*$ of a commutative algebra in spectra is a $\mathbb{Z}$-graded supercommutative algebra (Example 2.2 there).

On a different direction, there's also Kapranov's Supergeometry in mathematics and physics, which discusses in section 3 a conjectural relationship between supersymmetry and gradings by the sphere spectrum. See also the following MO question:

MO236266, What is the relation between the sphere spectrum and supersymmetry?

Incidentally, one specific realisation of this conjectural relationship is that if you take the sphere spectrum $\mathbb{S}$ and consider $\mathbb{S}$-graded rings in the sense of Bunke–Nikolaus (which turn out to be the same thing as rings graded by the 1-truncation $\tau_{\leq1}\mathbb{S}$ of the sphere spectrum), you obtain $\mathbb{Z}$-graded supercommutative rings back as a special case, although $\mathbb{S}$-graded rings are more general, and include also ordinary commutative rings as another special case, for instance. See

MO403355, Generalising supercommutativity as a grading by the $1$-truncated sphere spectrum

(a question I asked a while ago) for more details.
A: Superalgebras have been used in various questions of algebra in a very striking way. To give some instances:

*

*Kemer's proof of the fact that, over a field of zero characteristic, every system of identities expressible in terms of the product in associative algebra follows from finitely many of them (Finite basis property of identities of associative algebras).


*Nilpotence results by various authors, see, for instance, Zelmanov's celebrated proof of global nilpotence of Engel Lie algebras (Engelian Lie algebras), or the survey of Vaughan-Lee indicating some other directions (Superalgebras and Dimensions of Algebras).


*Shestakov's elegant examples and counterexamples in different varieties of nonassociative algebras (Superalgebras and counterexamples).
A: A very interesting connection between supergeometry and 'classical mathematics' is the Stolz-Teichner programme. They have shown deep connections between moduli spaces of supersymmetric quantum field theories and cohomology theories. In $0$d, we can recover de Rham cohomology and in $1$d we recover K-theory. In $2$d it is conjectured that we recover TMF.
A: An example that was motivating for me is the paper Vanishing Theorems for constructible Sheaves on Abelian Varieties by
Thomas Krämer and Rainer Weissauer
(https://arxiv.org/pdf/1111.4947.pdf) which uses a description of a certain category of complexes of sheaves as a Tannakian category, in particular as a category of representations of a super-group, to prove vanishing theorems for the cohomology of perverse sheaves on abelian varieties, which was in particular the missing step to describe a suitable category of perverse sheaves as a neutral Tannakian category, i.e. as a category of representations of a group.
So here, at least by this method, one needed to pass through the super case to get to the ordinary case.
A: In "Unimodality and Lie superalgebras" (citation below), Stanley uses the theory of Lie superalgebras to deduce some interesting combinatorial consequences about the unimodality of certain combinatorially defined
sequences of numbers, and the Sperner property for certain finite posets.
Stanley, Richard P., Unimodality and Lie superalgebras, Stud. Appl. Math. 72, 263-281 (1985). ZBL0614.17004.
A: In probability/statistical mechanics there are many interesting models that are best understood via equivalences with certain "superprobability" models that are easier to study.
The best known example is probably weakly self-avoiding walk: If we choose a path in the lattice of length $n$, weighted by $\exp(-\beta \cdot\#\{$self intersections}), what does such a path typically look like? When $\beta=\infty$ this is the usual (strictly) self-avoiding walk, which cannot visit the same vertex twice, whereas for finite $\beta$ this constraint is encouraged but not enforced.
An interesting conjectural feature of both this model and the usual strictly self-avoiding walk is that in sufficiently high dimensions the self-avoidance constraint is "only felt locally", and the self-avoiding walk looks much like an ordinary simple random walk on large scales (e.g. it still converges to Brownian motion under rescaling). The critical dimension for this simple ("mean-field") behaviour to kick in is believed to be $4$, and the model in four dimensions is expected to exhibit particularly delicate behaviour, with polylogarithmic corrections to the scaling of various interesting quantities compared to the mean-field predictions.
All the best rigorous results on this (mostly due to Bauerschmidt, Brydges, and Slade — see e.g. Critical two-point function of the 4-dimensional weakly self-avoiding walk) go through an equivalence between weakly self-avoiding walk and a certain supersymmetric field theory.
Recently a few other interesting combinatorial/probability models (including edge-weighted random spanning forests and the "vertex reinforced jump process") have been analysed via similar equivalences with supersymmetric field theories. A nice overview is given in Andrew Swan's PhD thesis "Superprobability on Graphs".
