There is an interesting case for agile development of infrastructure, applying lessons learned from the software world to civil engineering projects.
This is a useful metaphor, but there is an unfortunate limit to how far we can labor this point - software is abstract, in the sense that there's not a clear 1-1 translation between planning, design, forces that drive construction, and the actual physical output. Software is mostly non physical, and it works best when allowed to evolve so that the problems encountered and lessons learned from trying out an idea first flow into the ongoing development and evolution of a project. This concept is predicated on the assumption that mistakes will happen - and in the abstract realm of software, there is a lot more leeway for mistakes than in the process of concrete bridge building.
You can't push the idea of Don't Study It, Build It too far at a public policy level, before its utility becomes completely deflated. The problems are higher level than management and engineering. They are to do with decision making which takes place at a distance far removed from the actual costs and consequences of construction. Each prospective development is siphoned off into its own separate budget and study, encouraging supporters and detractors to take a black and white perspective of the complicated reality, where the context of multiple overlapping needs is forgotten.
Below all this lurks an even more awkward social and ethical issue. The origins of the word infrastructure itself are French, deriving from military construction, and it represents a militaristic, top-down approach to public works and community organization. Much of the human developed world is entirely reliant on action-at-a-distance style centralized public planning. Rather than recognize that as a species, we need to learn how to develop more efficiently, this command and control construction has encouraged many people to invent a separate category which they call "the environment", and refer to it in pseudo-Victorian splendor, as if it were a glorious and pure Other. This is why terms like unspoilt and pristine are so often used to describe wilderness and non-Urban landscapes.
It's time to revisit these assumptions. Because of a very dominant European historical influence, we make a weak distinction between arable land used for farming, urbanization, and wild ecosystems, but we could be completely wrong about many things we take for granted, such as the biodiversity of the Amazon Basin actually being the result of careful human cultivatation. If we can find any correlations to agile development in the future of place making, it is the idea of being iterative, cyclic, learning from our mistakes, through this can only happen when looking at human history on a long time-wave scale (though still blindingly quick on a geological time scale):
13,000 or more years ago people begin to kayak and fish along the edge of the Bering Land Bridge. They arrive in Alaska in several waves. A small genetically isolated group of individuals eventually making it past glacial shield walls into North America proper. Populations boom and through mismanagement quickly kill off all the mega-fauna. Large populations also act as a well minimizing any genetic drift from newcomers. Post glacial mega-floods wipe away planetary early record of civilization. Lacking oil based industrial technology they eventually come into a balance with their natural world, using a kind of bio-technics. Fire is used to burn off underbrush, to drive bison and to create grassland. Bison in turn are encouraged to constantly migrate (again by use of fire) and this helps the health of grasslands. The eastern seaboard is planted with fruits and nuts. A sublime expertise and attention to the pattern of the life emerges out of ongoing use of life technology. Even their language becomes deeply inflected with their naturalist lore. Strategies employed included an ambitious and possibly conscious genetics program to define maize. Populations grow up to about 100 million; comparable to current populations in many ways. Huge civilizations rise. With the arrival of Europeans populations crash due to genocide and disease; about 98% of the natives dying off within 100 years.
This story is remarkably similar to what occurred in Aotearoa over a slightly faster time scale. Like other human peoples, early Maori burned, cleared, hunted species to extinction, and eventually developed a cyclic pattern of existence based on conservation values imbued in a sophisticated social fabric. In any case, there are common human values of stewardship or custodianship (the Maori term for this is kaitiakitanga). We should be thinking the same way about environment and infrastructure instead of vacillating over the distinction.