Human-systems integration can advantageously use artificial intelligence to reach socio-technical flexibility. This paper proposes a systemic approach combined human and machine intelligence based on the shift from rigid automation to flexible autonomy. It emphasizes various issues, including maturity , life-critical systems (considering not only cognitive but also physical properties), systems as representations of people and machines, and the need to consider expertise and experience. It provides a deeper definition of the concept of system as a system of systems, represented by a network of systems where a network of functions could be dynamically allocated onto a structure of structures. The TOP model is introduced (technology, organizations and people). Three dual design and operations processes for human-machine function allocation are introduced: substitution/automation; amplification/interaction; and specula-tion/amplification. Three kinds of operational processes, leading to allocation of human and machine functions, are presented: procedure following that leads to automation of people; automation monitoring that results from automation of machines ; and problem solving that involves coordination of increasingly-autonomous people and machines. This is precisely where AI could be efficient and effective by supplying tools that augment people's capabilities in problem solving. The overall approach has been coined "FlexTech".