Since $1+ab$ is a unit in $R$, there exists an inverse $s$ such that
$$ (1+ab)s = s(1+ab) = 1. $$
Now let $t = 1 - bsa$. Then
\begin{align*}
t(1+ba)
&= 1 + ba - bsa - bsaba \\
&= 1 + ba - bs(1+ab)a \\
&= 1 + ba - ba \\
&= 1
\end{align*}
and likewise for $(1+ba)t$. So $1+ba$ is also a unit with the inverse $1-bsa$.
My intuition. Formally and heuristically, the inverse $s$ of $1+ab$ is
$$ s = (1+ab)^{-1} = 1 - ab + (ab)^2 - (ab)^3 + \cdots. $$
Likewise, if $1+ba$ has an inverse, it will have a formal expansion
\begin{align*}
(1+ba)^{-1}
&= 1 - ba + (ba)^2 - (ba)^3 + \cdots \\
&= 1 - b(1 - ab + (ab)^2 - \cdots)a \\
&= 1 - bsa.
\end{align*}
Then we can begin the actual proof with this guess.