In my experience, your ability to be believable when you say you can accomplish the work needs to be demonstrated as shown by your prior work as demonstrated by your publications in the area or your work done setting up the proposal. But otherwise your personal reputation as known by the people on the panel has less to do with it. The fit of your work to a program, however, is very important. The panel only gives a recommendation to the NSF Program Officer who will make a further recommendation to their boss which may go further up the chain. Those people will decide what projects get awarded under a program, and they have probably never heard of many of the awardees. The panel's recommendation will carry a lot of weight with the PO and their bosses, but it's not determinitive. Lower-ranked proposals can be elevated by the PO over the panel's recommendation and vice versa.
If the program is an ongoing or annual program, look at who has won prior versions of it using NSF's search page. If you know the luminaries in your field, expect to see some of their names, but also expect to see people you've never heard of. That's your clue that people with little or no reputation in the area are capable of winning awards in programs with a good fit. Read their abstracts, and see how well they align with the RFP/call.
There are very few awards in each program every year compared to the number of submissions. It varies from the high teens to low twenties across the directorates, and it's around 23% for the whole of the Foundation. The reasons for funding a particular proposal get pretty esoteric pretty quickly. I recommend finding the best fit, calling them, and asking them if you can give them a quick description to see if they think it's a fit, too. Most POs are happy to tell you if they think it's a good fit. They won't tell you if they'll fund it (they can't), but they can steer you around a bit. They may even steer you to a different program. Better to talk to them than a senior faculty member who has a hunch unless they sat on program A or B's panel last time around. If they did, they're not supposed to be telling you that.