I doubt that was the decision process. It’s much more likely that there is a directive coming down from the top that “we need to go all in on AI”, which then gets amplified down by middle management and the result is AI smeared over all results irregardless if helpful. That then drives up some vanity metric like “searches answered by AI summaries”, while metrics like “bad AI summaries shown” don’t get attention. As a result the organization is happy, people can get promoted, etc.