First off; to clear a couple of things up:
Furin cleavage sites in coronaviruses are quite common - about 30-50% of human ones have them naturally1, and these sites do NOT necessarily cause them to infect humans, it is just that we ubiquitously express furins2, so the presence of a furin-cleavable site makes them more likely to be capable of infection people, but also many other organisms too.
Gain-of-function (GoF) is used all the time in scientific research, as you can cause a gene to gain a function in a bacterium or mouse with almost no risk to anything outside of that individual organism. What you are asking about is GoF in a highly pathogenic virus that has the potential to cause damage to humans.
There are quite a few benefits to so-called GoF research
You could very easily argue that the AstraZeneca/Oxford vaccine, ChAdOx1 is a GoF for a chimpanzee adenovirus, as the chimpanzee virus will not normally infect humans, yet this virus has not only been engineered to infect humans, but also to carry the SARS-CoV-2 spike protein and no-one is complaining about this research at all
You can also use it to study historical viruses, such as the 1918 H1N1 influenza virus, where the sequences are typical influenza, but the only way to find out what properties of these gene products or which combinations to look out for, involve taking those genes and putting them into a lower pathogenicity virus and seeing whether they are more transmissible, pathogenic etc and what particular parts of the protein(s) make it so. We can then look out for these properties in regular surveillance and treat them as potentially more dangerous.
You can also use it to look at the pathogenicity of the viruses: what symptoms do they produce in a model organism, how fast does it act, how lethal are they etc. This is so that we know what to look for in medical terms and likely danger for people if you see a virus with a certain combination of features.
Thirdly - as pure research, if we know the features we are looking at, then finding out exactly what happens when these are introduced to a non-pathogenic virus is of scientific interest, it lets us know how and when viruses could become zoonotic. It also lets us do things like find out what happens when you take a virus like this and passage it through an organism (i.e. infect one, allow it to infect another, which then infects another and so on). This shows us how the viruses evolve, how they adapt to hosts, whether they become more pathogenic or less and many other things.
There's a nice write-up of the risks and benefits here3 for influenza, but the concepts can equally be applied to coronaviruses or any other virus.
1: Peacock Nature Microbiology volume 6, 899–909 (2021)
2: Braun Clinical and Translational Immunology (2019)
3: Casadevall mBio Vol. 5, No. 4 (2014)