Michel de Certeau's most influential conceptual pair divides the world of human action into two fundamentally different modes. A strategy is the operation of an institution, corporation, or system that controls its own territory—a lieu propre, a proper place—from which it surveys, calculates, and acts upon an exterior. A tactic is the operation of someone who does not possess such territory and must navigate spaces designed by others. The walker in the planned city, the cook working with ingredients at hand, the reader making unexpected meaning from a text—all are tacticians. The distinction is not about power versus weakness, but about two qualitatively different relationships to systems: one that designs them, one that inhabits them. AI platforms are strategic territories; builders navigating them are tactical practitioners.
De Certeau developed this framework through close study of everyday life: walking in cities, cooking in kitchens, reading texts, navigating bureaucracies. He observed that the Western intellectual tradition had systematically ignored the creativity of people who did not control the systems they inhabited. The genius in the studio, the author at the desk—these figures possessed a lieu propre and operated strategically. But most human creativity, de Certeau argued, is tactical: it works within constraints imposed by others, seizing opportunities without controlling the terrain. The strategic view from above—the urban planner looking at the grid—sees coherence, order, legibility. The tactical view from below—the pedestrian choosing the shortcut through the alley—sees something else entirely: the city as practiced, lived, transformed by use into something the designer never intended.
The AI platform exemplifies strategic territory. A large language model is trained on a corpus selected by its builders, structured by architectural decisions, governed by alignment constraints, deployed under commercial terms. Every user enters this territory without having designed it. She prompts the model, receives output shaped by decisions she did not make, and works within a space whose rules were set by engineers, ethicists, and executives she will never meet. Yet the user is not passive. She navigates. She selects useful output and discards generic fluff. She combines multiple responses, inflects the model's prose with her own voice, refuses the confident error. These operations are tactical—opportunistic, moment-seizing, creative engagements with a system that was not designed for her specific need but that she makes serve her purpose anyway.
The distinction maps onto the experience every AI-assisted builder recognizes: the platform controls the model, sets the terms, can update overnight in ways that break yesterday's workflow. The builder cannot change this. But she can walk differently through the territory—can develop a practice that finds gaps the designers overlooked, capabilities the documentation does not mention, combinations the strategic logic did not anticipate. Prompt engineering is tactical art. Bricolage with AI output is tactical creation. The freedom available to the builder is not the freedom to redesign the model. It is the freedom to practice—to engage with the model's output in ways that transform strategic production into tactical meaning.
The strategy-tactic distinction emerged from de Certeau's early 1970s fieldwork in urban sociology and cultural analysis, crystallizing in his 1974 Culture in the Plural and reaching definitive formulation in The Practice of Everyday Life (1980). He drew on Clausewitz's military theory, Foucault's analysis of disciplinary institutions, and his own observations of how ordinary people navigate Parisian housing bureaucracies, public transit, and commercial systems. The framework synthesized his Jesuit theological training—attentive to the relationship between institutional structure and individual spiritual practice—with structuralist and post-structuralist currents in French theory. De Certeau's originality lay in refusing the structuralist reduction of individuals to passive bearers of cultural codes while also refusing the liberal fantasy of the autonomous subject. His practitioners are neither determined nor sovereign—they are creative within constraints.
Territory defines the difference. Strategies require a lieu propre—a place of one's own from which to generate relations with an exterior. Tactics operate in the territory of the other, without the spatial or institutional base for strategic planning.
The walker writes the city. De Certeau's paradigmatic case: the pedestrian who takes the urban planner's grid and transforms it through her walking into a personal, lived geography the map cannot capture.
Tactics are opportunistic. The tactician has no long-term plan, no comprehensive vision—only the capacity to seize the moment, exploit the opening, make use of whatever the strategic system provides.
AI platforms are strategic; builders are tactical. The model defines output space. The builder navigates it, finding paths the designers did not foresee—prompt engineering, creative misuse, selective appropriation of generated material.
Tactical freedom is real but conditioned. The practitioner cannot redesign the system, but she can transform what the system produces into something genuinely hers through the quality of her engagement.
Critics argue de Certeau romanticizes the weak, treating everyday evasions as resistance when they leave power structures untouched. Scholars like Lila Abu-Lughod question whether 'resistance' retains meaning if it describes every deviation from institutional norms. The framework's application to AI raises the question whether tactical navigation of a platform constitutes genuine creative agency or merely operates within bounds the platform's business model already anticipates and exploits.