If you hire an all-female sales staff, and each employee is technically better in every way than average men, but your customers simply do not want to be sold to by women, then it doesn't matter how good the women are, how many extra skills they have that aren't priced into the market, you will fail.
That's the thing with systematic bias: it's systematic.
That all said, there is still a way to capture the moneyball-style giant pile of cash (and yes, I do believe it is there. There's one for black people too, and every group we think is less employable for certain jobs than white men):
You have to create a feminist company and a feminist market and spin them up simultaneously.
The reason this doesn't happen all the time is it's much harder to create a 2-sided market than a 1-sided one.
And frankly, due to the systemic nature of sexism, a 2-sided market probably isnt enought. You really need to create an n-sided market. You need suppliers and partners and sister companies and clients all on the same page, at least to the extent that your interface with those organizations is human-rich enough to permit sexism.
This is the same reason why anarchist (property-free) businesses have been hard to create, even though the fundamentals should be more efficient than a capitalist company. For it to work you need to spin up n-anarchist companies at once so they can feed off each other. Instead people try to create them one by one, so they fail. They are chewed up by the fundamentally antagonistic world around them... Same as pro-women organizations are.
It's a hard startup problem, but we are getting good at solving those. I wouldn't bet against this being solved within 20 years, at least in proof of concept.
I don't know that you can blame anything on "the fundamentally antagonistic world" since all businesses face a fundamentally antagonistic world and their goal is to overcome that.
One could certainly imagine a law firm matching their hiring practices to the prejudices of their clients, and having this actually be the rationally optimal strategy, i.e. the impact on customer relationships is larger than the inefficiency of rejecting many otherwise good candidates.
It would make sense if your lawyers roughly mirrored the demographics of your customers - having the same social values, regions, schools, hobbies, accents as your customers do. And customers of the expensive law firms are quite different from the general population.
Take software engineering instead as an example, you should be able to hire an all women team of top engineers by paying regular market salaries. There would be no downside and all upside if the gap between male and females of the same skill level is true.