Evolutionary processes largely rely on the production of diversity. Genetic robustness, by allowing the accumulation of neutral diversity within a population, has been associated with increase in evolutionary potential (evolvability). We have already shown that one can use a well-known source of robustness, the redundancy of the genetic code, to alter the evolvability of any protein. The topology of the code allows synonymous codons to sample different mutational neighborhoods. Using this property, we developed an algorithm to design synonymous sequences with maximally divergent evolutionary potentials relative to the input sequences (see publication). A web version is accessible here for academic use. At the population level, each of these sequences expands the scope of the evolutionary landscape that can be explored by the encoded protein, and ultimately increase the odds of uncovering adaptive mutants. We applied this principle to evolve new antibiotic resistance phenotype variants. Fundamentally, our results provide an example of how neutral diversity may favor evolvability. Moreover, in light of the rapid development in nucleic acid synthesis, the use of rationally designed synonymous genes offers a profitable enhancement to any directed evolution procedure. We are currently applying this principle to other genes of interest, in order to access to novel phenotypes, so far undescribed.