First, Unconditional Cash Transfers (UCT) work better than most people expect (http://www.economist.com/news/international/21588385-giving-...). As noted in the article, UCTs work surprisingly well in scenarios where lack of capital is the primary problem. A UBI (at least as usually is discussed in the US) is a de-scaling UCT.
Second, most proponents of a UBI see it as more efficient because of the savings on administrative overhead. So dollar for dollar the assumption is that your social safety net with UBI is larger than current welfare.
Which finally brings us to the observation that current welfare is not an all-inclusive safety net, so opposing UBI because it isn't either seems a bit of a distraction.
A more interesting question, raised by the Economist article linked above, is where do UCTs break down? This is a question whether or not UBI should be _entirely_ UCT or if there should also be some Conditional Cash Transfers (CCT). This would be like providing incentives for going to college.
CCTs are better at correcting deeper issues that contribute to the cycle of poverty.
Ultimately I feel like a UBI is a good first step, but ultimately a combination of UCTs and CCTs are needed to really combat poverty.