Link to the article: https://pubmed.ncbi.nlm.nih.gov/32584464/
They reviewed 57 papers looking at the sensitivity and specificity of their testing. They found that testing was useful early on but was useful after >15 days. After 35 days, data is limited and so using it determine prevalence in the general population is limited. Data available is limited and selective and more transparency is needed for a better evaluation.
TL;DR The utility of antibody testing in the beginning of an infection is low but appears to be useful later on (15-35 days after symptom onset)