McKinsey & Company's Ignacio Gorupicz, a leader of the Service Operations practice in Latin America, stated on 2026-09-03 that Agentic AI forces Global Business Services (GBS) to evolve beyond traditional capabilities of scale, standardization, and cost reduction. ^[1] Gartner estimates that businesses will allocate $2.59 trillion to AI spending in 2026. ^[2] Jerry Haywood, CEO of boost.ai, stated that businesses are not seeing AI ROI because they have not dismantled their processes and remade their organizational workflows to harness AI's power. ^[3] McKinsey & Company calls this the 70-20-10 rule: 70% of the business value from an AI program comes from people, culture and process change; 20% comes from data infrastructure; and a mere 10% stems from the actual AI algorithms. ^[4] Several high-profile studies, including a widely discussed MIT study, found that most organizations are not seeing a return on investment from their AI spending. ^[5] According to INEGI data, micro, small, and medium economic units concentrated 70.7% of employed personnel and generated 56.5% of the income of economic units. ^[6] Due to high operational costs, large banks often prioritize high-volume segments like consumer or corporate loans, leaving micro, small, and medium enterprises (SMEs) as a lower priority. ^[7] McKinsey & Company reports that performing operations via application costs 0.03 USD, which is 95% cheaper than the 0.65 USD cost of branch transactions. ^[8] Traditional banks historically rely on physical networks, which incur significant operational costs including rent, cash handling, maintenance, and on-site staff. ^[9] According to the analysis, Agentic AI adoption could reduce the size of Global Business Services teams by 20% to 30% while simultaneously making them significantly more productive. ^[10] The article asserts that while traditional automation executed predictable tasks, Agentic AI can incorporate judgment, synthesize information from different contexts, and participate in decisions previously requiring human intervention. ^[11] The author argues that as AI assumes more transactional work, traditionally administrative functions must evolve toward advisory, strategic, and decision-making roles. ^[12]