Zika doesn't alter the host genome at all.
Since Zika can alter the genome of a baby to cause deformities
Zika virus is incapable of altering the host genome. The exact mechanism by which Zika virus infection causes microcephaly is still unclear, but basically it depletes the neural stem cells at a point when the brain is developing, meaning brain development is more limited (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5516183/).
Some viruses do alter the host genome and are candidate technologies for gene therapy
Retroviruses are RNA viruses that copy their genetic material into DNA which is then integrated into a host genome. As such, and as you note, this can potentially be exploited to deliver new genetic material into an organism's genome. I'm not sure what you meant by transduction above, but the term is 'retroviral vector-mediated gene delivery'.
Retroviruses have been trialled for gene therapy for decades. See this sad report on a 2002 trial that apparently succeeded but had unexpected side effects here. However, CRISPR is probably an easier method to use.
You wouldn't attempt to modify a fetus
Some terminology here: a fetus is what an embryo grows into; it's getting bigger and more complex. And as illustrated by Zika-related microcephaly, any messing about with fetal development can have nasty side effects. There are also more cells to modify and you're likely to end up with a 'mosaic' where some cells are genetically different to others within the same organism. If you're attempting to modify the somatic cells of one organ, you might as well do it after the baby is born; if you're attempting to modify every cell in a human (including germ line cells) you'd do it to an egg or embryo.
This study describes a successful attempt to modify embryos using CRISPR. They saw only 1 mosaic in 58 embryos.