PMFJI, as jennyp likes to say. Mutations happen. Many are harmful, but then many are neutral and some are beneficial. A sexual species population is a cloud of unique genomes with a host of mutations floating around in it. Natural selection prunes what doesn't work.
A common question is where new information comes from. The genome enlarges during a certain kind of copy error called a duplication mutation, which creates an extra copy of a section of the original DNA. This can include one or more genes or even the entire genome. Having two copies of an essential gene frees one up to mutate and possibly produce a protein similar to that produced by the original gene but useful in some new way. Hemogolobin, the molecule that carries oxygen in blood, apparently originated long ago in some simple organism from myoglobin, which does a similar function in muscles.
So mutations can be beneficial, can enlarge the genome without being harmful, and can eventually produce new functions. Where's the inadequacy?