Too late to take corrective action?

Develop a workaround if possible; and if that fails, then yes populate the lessons learned to avoid such a mistake in the future.

I was called one time by a friend to visit his own house site (2000 sqm land) in Oman, and I was curious about the foundations piles, and I found that the whole thing was done wrong. We called in to site the PM ,surveyer, consultant, and piling company. Thanks god that I was not in the pm shoes at that day.
It was very messy, the project was late at least 2 months.
I am sure the consultant did not visit the site ever. PM visited the site frequently but he took the work of the piling company without inspection and obviously he is one of those who doesn't roll his sleeves. Surveyer marks were a bit confusing. When we asked for the inspection plan, forget about it. Etc...

Mistakes happen. All the time. To err, after all, is human.

As to how to deal with it, the other responses have covered it pretty comprehensively already.

The only qualifier I would stress on is to first contain the fallout instead of looking to assign responsibility (or 'blame') for the mistake.

There's a story I heard as a kid. A man and his wife are walking when a snake comes up and bites the wife. The snake slithers away and the husband, full of rage, chases after it. He catches it after about a 100m and stomps it to death.

But when he comes back, the wife has succumbed to the poison.

Moral of the story: Get the wife medical care, first. The snake can always be hunted down later - and even if it is not, the wife is still alive.

Good story Karan. Unless he really didn't like his wife (analogy: wants the project to fail and assign blame which happens a lot).

1st Point: what means too late to do corrective actions? Many proposals above still recommend corrective actions. So your delivery does not match with the agreed scope, that was proposed to the customer. Possibility no 1 check how critical the failed criteria is for your customer. Is there a chance to compensate the customer for changing the criteria? Approach no. two: check your legal exponation, what is the worst case scenario (e.g. safety relevance, see above Mr. Conte) and prepare your organization. Approach no 3: have some idea in place how to help your customer out, BEFORE you talk to him.
My general comment. You are not alone. The bigger projects are the more things happen like this, always try not to be the most painful supplier to your customer as those always get special treatment :-).

Its belongs, at least in my personal experience, to quality. Then it is a matter to make a difference between quality and grade. Sometimes what you stated belongs to grade. Both have to be defined at the begining of the initiative and belongs to strategical decisions that are beyond the initiative. The second is what it does mean "to late to take corrective actions"? Always is time to take actions. Third, and at the end, is a matter of risk. The risk must be evaluated. For example, if the risk has as impact in loosing human being life then you have to consider to put the corrective action in place (I can list here well known and documented situations where that did not happen). As any other thing inside the project is not a decision that a project manager must take.

Mistakes happen. All the time. To err, after all, is human.

As to how to deal with it, the other responses have covered it pretty comprehensively already.

The only qualifier I would stress on is to first contain the fallout instead of looking to assign responsibility (or 'blame') for the mistake.

There's a story I heard as a kid. A man and his wife are walking when a snake comes up and bites the wife. The snake slithers away and the husband, full of rage, chases after it. He catches it after about a 100m and stomps it to death.

But when he comes back, the wife has succumbed to the poison.

Moral of the story: Get the wife medical care, first. The snake can always be hunted down later - and even if it is not, the wife is still alive.