Both the Atiyah-Hirzebruch and the Serre spectral sequence can be constructed from a skeletal filtration of a CW-complex:
We can construct the Atiyah-Hirzebruch spectral sequence by filtering $X$ by its skeleta and applying an exotic cohomology theory $h$. This gives an exact couple, and the spectral sequence associated to this runs $H^p(X; h^q(pt)) \Rightarrow h^{p+q}(X)$.
Given a fibration $E\rightarrow B$ with typical fiber $F$, we can filter the base $B$ by skeleta. This induces (by taking preimages) a filtration on $E$, and under mild assumptions, we arrive at the Serre spectral sequence running $H^p(B; H^q(F))\Rightarrow H^{p+q}(E)$ (by applying ordinary cohomology to the filtration).
Now we can combine those two constructions into one, obtaining a spectral sequence running $H^p(B; h^q(F))\Rightarrow h^{p+q}(E)$. This specializes to the Serre spectral sequence by setting $h$ to be $H$, and to the Atiyah-Hirzebruch spectral sequence by considering the fibration $X \rightarrow X$, where $F$ is just a point.
(This is pointed out, for example, in Hatcher's "Spectral sequences in Algebraic Topology", where he mentions this as possible construction of the AHSS)
I have never seen any application where this tells you substantially more than, for example, first computing $H^{p+q}(E)$ via the ordinary Serre SS, and then $h^{p+q}$ via the ordinary AHSS. I do know some examples where the answer is slightly different from what you'd expect at first, for example the Hopf fibration $S^3\rightarrow S^2$ and $h^*=KO^*$. However, in these examples it seems to me that obtaining the actual structure of the spectral sequence requires considerable work with the ordinary ones.
Does anyone know of a striking example where having the generalized spectral sequence tells you more than just the two special cases?
