Figure 2
The diagram is divided into two sections. The top section is titled “A I adoption in S and O P” with an arrow pointing down to four main sections labeled “Performing paradox,” “Organizing paradox,” “Belonging paradox,” and “Learning paradox.” Each main section is outlined with dashed boxes and contains smaller labeled boxes arranged in columns. Under “Performing paradox,” the top row contains: “Attaining short-term efficiency with existing tools,” (left). “Reactive (fire fighting) demand-supply balancing” (right). Below in the middle, a box has the text “Relying on gut-feeling decisions.” At the bottom, two boxes, arranged horizontally, are labeled “Awaiting fully developed A I maturity,” (left) and “Proactive (A I-driven) demand-supply balancing” (right). At the bottom center, a box says “Indisputably following AI-generated insights.” A vertical two-way arrow labeled “S P 1” connects the top left box to the bottom left box, and a vertical two-way arrow labeled “S P 3” connects the top right box to the bottom right box.” A vertical two-way arrow labeled “S P 2” connects the top center box to the bottom center box. Under “Organizing paradox,” the top row contains: “Centralized A I implementation” (left) and “Relying on data silos” (right). In the middle, the box says “Maintaining traditional S and O P processes.” At the bottom, two boxes arranged horizontally are labeled “Decentralized A I implementation” (left) and “Integrating A I-driven data systems” (right). Below these, centered at the bottom, is a box labeled “Redesigning to A I-based S and O P frameworks.” A vertical two-way arrow labeled “S P 4” connects the top left box to the bottom left box, and a vertical two-way arrow labeled “S P 6” connects the top right box to the bottom right box. A vertical two-way arrow labeled “S P 5” connects the top center box to the bottom center box, Under “Belonging paradox,” the top row contains: “Collective A I-driven S and O P identity” (left) and “Resisting A I-driven S and O P” (right). Below in the middle, a box reads “Streamlining A I-driven collaboration.” At the bottom, two boxes are arranged horizontally: “Individualistic S and O P identities” (left) and “Embracing A I-driven S and O P” (right). Below these, centered at the bottom, is a box labeled “Preserving specialized S and O P expertise.” A vertical two-way arrow labeled “S P 7” connects the top left box to the bottom left box, and a vertical two-way arrow labeled “S P 9” connects the top right box to the bottom right box. A vertical two-way arrow labeled “S P 8” connects the top center box to the center bottom box. Under “Learning paradox,” the top row contains: “Developing traditional S and O P skills” (left) and “Making incremental improvements” (right). In the middle, a box has the text “Relying on human learning.” At the bottom, two boxes are arranged horizontally: “Acquiring A I-related competencies” (left) and “Embracing disruptive A I innovation” (right). At the bottom center, a box is labeled “Leveraging machine learning.” A vertical two-way arrow labeled “S P 10” connects the top left to the bottom left box, and a vertical two-way arrow labeled “S P 12” connects the top right to the bottom right box. A vertical two-way arrow labeled “S P 11” connects the center top box to the center bottom box. A line drawing of a person with a round head, seated at a laptop, is shown centered below the four boxes. The text below the figure reads “S and O P professional.” Curved arrows extend from either side of the S and O P professional, and lead to “Performing paradox” on the left and “Learning paradox” on the right. The bottom half of the figure is divided into vertical columns under the heading “Impact pathways.” Each numbered column gives “IP1: Balancing immediate S and O P actions with long-term AI-driven strategies,” “IP2: Navigating AI in S and O P via centralized systems, process redesign, and data unification,” “IP3: Harmonizing AI-driven S and O P identities, collaboration, and technology acceptance,” “IP4: Bridging traditional S and O P human skills with innovative AI competencies.” The last row is “IP5: Managing the interrelated paradoxes of AI adoption in S and O P.” In the bottom section, four columns are shown. Each column is divided into three vertical sections separated by horizontal dashed lines. In the first column, the first section reads “ The bottom section shows four horizontally arranged columns under the heading “Impact pathways.” Each column is divided into three vertical sections separated by horizontal dashed lines. The first column is labeled “I P1: Balancing immediate S and O P actions with long-term AI-driven strategies.” The second section reads “R Q: How can S and O P teams effectively balance the use of existing tools with A I implementation to optimize both short-term performance and long-term strategic benefits?” and “R Q: How can firms leverage ‘shadow runs’ and parallel implementations to develop A I maturity in S and O P while maintaining operational continuity?” The third section reads “P I: Gradually implement A I in S and O P through small-scale pilots (‘shadow runs’) without disrupting current operations,” “P I: Enhance A I models by incorporating S and O P domain expertise (‘tribal knowledge’) through structured feedback loops,” and “P I: Identify common disruptions from historical data and feed them into AI models to proactively reduce firefighting tendencies.” The second column is labeled “I P 2: Navigating A I in S and O P via centralized systems, process redesign, and data unification.” The second section “R Q: How can A I systems be designed to balance centralized control with local adaptability in diverse S and O P environments question mark” and “R Q: How can firms implement unified data governance while preserving function-specific data autonomy across S and O P operations?” The third section reads: “P I: Develop centralized A I governance to secure data access while allowing model customization for regional S and O P needs,” “P I: Integrate A I with traditional S and O P processes to enhance workflows and improve model accuracy through continuous training,” and “P I: Establish a centralized data layer as the organization’s ‘single source of truth’ while supporting local access needs for specific S and O P functions.” The third column is labeled “I P 3: Harmonizing A I-driven S and O P identities, collaboration, and technology acceptance.” The second section reads: “R Q: How can firms develop A I systems that promote standardization while accommodating diverse organizational cultures and functional needs in S and O P?” and “R Q: What mechanisms enable A I models to effectively embody different functional perspectives while fostering cross-functional S and O P integration?” The third section reads: “P I: Train A I models to ‘think like’ each S and O P function while aligning with specific departmental needs,” “P I: Leverage A I to enhance collaboration across S and O P functions by integrating diverse expertise for more informed decisions,” and “P I: Automate routine S and O P tasks to increase efficiency and cut resource costs while supporting employees in transitioning to new A I-driven roles.” The fourth column is labeled “I P 4: Bridging traditional S and O P human skills with innovative A I competencies.” The second section reads: “R Q: What are the most effective strategies for upskilling S and O P professionals to bridge the gap between traditional expertise and AI-driven competencies question mark” and “How can firms develop S and O P professionals’ ability to configure and direct AI systems while preserving their essential domain expertise question mark” The third section reads: “P I: Train S and O P professionals to set operational boundaries for A I models (‘guardrails’) without requiring advanced technical skills,” “P I: Preserve human input for creativity, negotiation, and complex decisions to balance A I efficiencies with human judgment,” and “P I: Apply incremental A I adoption strategies that gradually build accuracy and trust without overwhelming current workflows.” Below these columns, a horizontal section is divided into three sections. From top to bottom, the text in each section is “I P 5: Managing the interrelated paradoxes of A I adoption in S and O P.” The section reads: “R Q: How do the various paradoxes in A I adoption for S and O P interact, and what strategies can firms employ to manage these interrelations effectively question mark” and “R Q: How do governance structures and process redesigns co-evolve during A I implementation across interconnected S and O P functions?” The third section reads: “Embrace a systems-thinking approach to anticipate and manage A I’s ripple effects on various S and O P functions to ensure a cohesive transition,” “P I: Implement strategies that consider interdependencies between A I-driven and traditional S and O P workflows for balanced integration,” and “P I: Use change management principles to address multiple tensions and maintain stability as AI becomes integrated into S and O P processes.” Below the table, definitions for the terms are given as “S P: Sub-paradox, I P: Impact pathway, R Q: Research question, P I: Practical implication.”Revealed SPs and proposed pathways