In the Wikipedia article, it seems to me as if it's implied that it is enough to check the conjecture only for maximal gaps (numbers $n$ s.t. $\forall k<n:g_n>g_k$).
I.e it holds that $P_{k+1}<P_k^{1+\frac{1}{k}}\Rightarrow P_{m+1}<P_m^{1+\frac{1}{m}}$ where $P_{k+1}-P_{k}$ is a maximal gap and $m<k$. (Or some similar statement that involves maximal gaps.)
Is this really the case or perhaps I misread the Wikipedia article? I tried to prove it using high school manipulations, but without much success.
Define $q(x)$ to be the smallest prime greater than $x$ and $$ f(p)=\begin{cases}1&\text{, }p\text{ is prime and }q(p)<p^{1+1/\pi(p)}\\ 0&\text{, otherwise}\end{cases} $$
Then Firoozbakht's conjecture is that $f(p)=1$ for all primes $p$. You seem to be asking if $f(p)=1$ for all primes $p$ beginning a maximal gap implies Firoozbakht's conjecture, and it does.
I should also mention that Firoozbakht's conjecture is believed to be false. It is expected that for any $k<2/e^\gamma=1.1229\ldots$ there are infinitely many prime gaps of length greater than $kp\log^2p$ and Firoozbakht's conjecture fails if there is even one with $k\ge1.$ It was a reasonable conjecture in 1982 when it was first posed but advances in the mid-80s and later have convinced number theorists that it's highly unlikely.