ICSI is particularly helpful for patients who suffer from male factor infertility, lowering the likelihood a cycle will be cancelled due to failed fertilization.
For many male factor patients, using ICSI is a must, especially in circumstances of azoospermia. Where ICSI's use is more controversial is in the context of poor morphology, but an otherwise normal sperm count and rate of motility.
Where ICSI is generally not necessary is in cases where there is not a history of male factor infertility, with the glaring exception being if a couple's eggs did not fertilize in a previous cycle.
There are numerous circumstances where doctors are inclined to use ICSI "to be on the safe side" and, in those cases, ICSI rarely, if ever, helps.
One of the scenerios where doctors want to deploy ICSI is in the case of PGS testing. There is debate if ICSI should be a requirement when PGS is used, and there is no data yet showing that IVF cycles with PGS-tested embryos do better when ICSI was performed than when it wasn't.
The financial costs of ICSI may appear trivial in the context of larger IVF expenses, but they're still ultimately meaningful, coming in at $1,000 - $3,000 depending on the city and clinic where you're being treated.
ICSI is a micro-surgery, and a delicate one at that. Roughly 5% of eggs are damaged in the process and the long-term effects are still being studied. In larger studies in Israel and Australia, ICSI has been associated with higher rates of birth defects, but it's hard to disentangle whether this is due to the underlying condition which prompted its use in the first place.