Pinal Dave wrote up his methodology for what to do when you discover bad data in production. If you’ve never been in this situation, having a plan beforehand is critical. Likewise, if you’ve been in this situation but it went badly, this is a good high-level approach. Compare what happened in your case with what Pinal recommends.
For instance, his first step is triage. People/teams without a proper plan often dive in and start trying to fix stuff. That’s the wrong approach because you don’t know the scope of the problem. The immediate fixes could lead to more problems. And he concludes with the blameless review. If blame is going to be attached to a postmortem, people are going to be guarded and you won’t get a proper root cause analysis or it will come with increased difficulty.