Generative AI (genAI) is spreading rapidly across organisations, yet many firms remain stuck between experimentation and stalled integration. Drawing on qualitative interviews with managers in 18 Swiss organisations, this exploratory, inductive study shifts attention from early adoption to implementation-for-integration of genAI. We show how genAI’s general-purpose character, its role within hybrid intelligence configurations and its visibility at the brand-stakeholder interface stretch classical adoption and acceptance models beyond their usual scope. The analysis yields a five-dimensional framework in which implementation emerges as a non-sequential, interdependent configuration rather than a linear rollout. Sustained integration appears to depend less on isolated genAI assets and more on dynamic capabilities that keep the organisational rationale for genAI alive over time. These capabilities repeatedly reconfigure how genAI is used and governed as its boundaries evolve. Theoretically, the study recontextualises classical adoption perspectives for a disruptive and transformative technology and extends them towards an integration-focused implementation logic. In doing so, the framework illustrates conditions associated with moving from scattered experimentation to more stable, value-creating genAI use, inviting a potential redefinition of marketing’s role at the brand-stakeholder interface. Generative AI (genAI) now fuels bold ‘AI-first’ declarations across industries. Yet many firms end up with largely symbolic moves and, in some cases, public U-turns when these ambitions trigger negative stakeholder reactance at the brand-stakeholder interface. Drawing on interviews with managers in 18 Swiss organisations, this chapter identifies conditions that appear particularly important for moving from early experimentation towards integration. Because genAI is a general-purpose technology, embedded in hybrid human-AI work and directly visible at the brand-stakeholder interface, progress often depends on coordinated change rather than isolated pilots. We propose a five-dimensional framework for implementation: (1) Response to External Factors, (2) Culture & Individuals, (3) Resources, (4) Activities, Processes, Outcomes and (5) Structures & Governance. This framework highlights conditions associated with more stable, value-creating use of genAI. While AI-first statements can draw attention, the chapter highlights the quieter organisational work that fosters sustained integration. This includes balancing initial efforts (e.g. communication, fostering psychological safety, training) with deeper shifts in governance, processes, revenue models and external relationships as genAI’s role expands. Rather than offering a single blueprint, the framework is intended as a lens to help managers reflect on how these issues play out in their own organisations.