Yes of course. Did you? The balance of evidence is strongly toward polygraphs having non-trivial predictive power. People frequently conflate unreliability with being fully random.
If you require applicants to pass five independent tests that have 60% accuracy (e.g., interview, polygraph, family/friend interviews, recommendations, criminal record), good applicants are more than 7 times more likely to pass than bad applicants.
Indeed, most of our important decisions are made by combining many weak pieces of evidence. Even if some other tests are more accurate, additional less-accurate tests can easily be worth running if their costs are modest.
Again, this is not an argument that the information provided by the polygraph actually passes a cost-benefit analysis in any particular case, it's just a criticism of the "all or nothing" mindset regarding accuracy that many folks have on this topic.
