Many desirable chemical transformations have yet to be discovered. Although most transformations are achievable if constraints such as the number of manipulations, overall yield and available resource are removed, there is a growing realisation that, with a relatively small set of new bond-forming reactions, target-oriented synthesis would be transformed. The new reactions would need to be complexity-building, using functionality in newer and smarter ways. As indicated above, they would need to be clean, atom efficient and selective so as to be used in sequence with other catalytic and reagentless transformations without need for protection. Examples include reductive conversion of a secondary alcohol into a carbon nucleophile with retention of configuration; oxidative coupling of unfunctionalised sp3-hybridised carbon atoms; internal redox reactions that relay reactive centres around a molecule as desired, etc. New ideas for identifying transformative reactions need development (e.g. examining impact by adding ‘unknown’ reactions in computational retrosynthetic analysis programmes).