Nearly nine out of ten organizations regularly use AI today, yet most have failed to realize material enterprise-level benefits. It is becoming clear that adoption isn’t just happening slowly, it’s happening unevenly. Organizations don’t begin their AI journeys from the same starting line. Companies vary dramatically in their data readiness, organizational and process structure, and cultural capacity for transformation. These differences create significant disparity in how quickly organizations can adopt AI, regardless of their aspirations.
This organizational unevenness is combined with the strategic and operational complexities of transforming different types of work at different speeds. Across service operations, spanning customer experience, middle office processes, and back office functions, AI adoption is uneven because workflows range from highly structured, rules-based tasks to those requiring significant human judgment, and each transforms at its own pace.
The types of work that are amenable to early success with AI implementation include rules-based interactions in call centers and back-office functions (i.e., accounts receivable, accounts payable, and contract matching). These structured environments benefit from clear processes and centralized control points, allowing for rapid AI deployment. In contrast, work that has historically resisted automation offers substantial AI potential but demands more complex implementation.
The challenge intensifies when work is distributed rather than co-located. When operations are centralized, the primary change point is a system that everyone uses; when distributed, the change point becomes individual human behavior across multiple locations. This fundamental difference explains why AI moves quickly through structured areas but slows dramatically in human decision-making domains where change management becomes paramount.
Why progress is stalling
What’s happening in most organizations isn’t deliberate, scaled AI implementation, but rather organic experimentation without clear coordination or value focus. This approach dilutes impact and makes demonstrating meaningful returns nearly impossible. McKinsey data shows that while 88% of organizations deploy AI in at least parts of their operations, almost as many report no significant bottom-line impact.
A critical mistake involves spreading focus and investment too thin, rather than identifying where AI creates the most value and concentrating resources accordingly. Each enterprise has a set of economic leverage points that create disproportionate value when AI is applied, yet many enterprises are applying AI everywhere without a clear linkage to value. This unfocused approach dilutes impact and makes it difficult to demonstrate meaningful returns on AI investments.
Many companies treat all experience-driven work and processes identically, applying a one-size-fits-all approach. Workflows ranging from highly structured to judgment-intensive require differentiated strategies, yet organizations often deploy AI uniformly and hope results will follow. This approach ignores the reality that while AI advances rapidly in structured areas, it encounters significant friction in domains requiring human judgment and adoption.
Organizations also struggle with unrealistic expectations, loading AI deployments with requirements far more stringent than those applied to human workers, then assessing performance against these inflated standards. Positive outcomes are more likely when entire workflows and processes are reimagined to take the agents and people into account and when the development of the agent is considered as intentionally as the development of the employee. In addition, many leading companies take a “fresh start” approach, designing the workflows without constraints of existing organizational inefficiencies
The Change Management Imperative
Even when AI tools are deployed in a targeted strategy, a critical gap emerges between deployment and adoption that organizations consistently underestimate. The real challenge is often in the change management required. A well designed agent workflow means little when the people it serves don’t trust it, understand it, or see sufficient reason to change established working methods.
The investment required to bridge this gap is substantial and often undervalued. Research shows that for every dollar spent on technology, organizations need to invest roughly two dollars in change management, capability building, and adoption to fully realize the benefits. This 2:1 ratio underscores a fundamental truth: technology deployment represents only one-third of the transformation equation, while the human and organizational elements comprise the remaining two-thirds.
As the nature of work changes, there will inevitably be changes to how humans work. The complexity multiplies when dealing with distributed workforces where standardization proves elusive. Disparate processes across cost centers, from personal spreadsheets to informal workarounds, require standardization before AI can be effectively applied. McKinsey’s research reveals that 86% of leaders feel their organizations are not very prepared to adopt AI in day-to-day operations, highlighting the readiness gap.
Infrastructure and Organizational Foundations
Unlike manufacturing or financial services, the services sector historically lacked an ERP or any other comprehensive technology ecosystem. This absence of foundational digital infrastructure creates compounding challenges when attempting to implement AI at scale, requiring organizations to build both the infrastructure and the AI capabilities simultaneously.
More fundamentally, organizations need to build the institutional structures needed to create, optimize, and orchestrate a new hybrid workforce. The McKinsey State of AI 2025 report finds that agentic AI proliferation is already outpacing the governance structures organizations have in place to oversee it. Data and technology foundations remain fragmented in most organizations. Different AI pilots operate in silos using disparate tools and infrastructures, making it difficult to reuse capabilities, standardize security measures, and maintain pace with rapid model updates. Without unified architecture, companies face challenges scaling successful pilots across the organization.
The Road Ahead
Appreciating the different speeds of AI readiness for service operations presents an opportunity for leaders to be strategic in their AI investment. Leaders who understand the complexity of transforming work that spans a vast spectrum from structured to judgment-based, co-located to distributed, deterministic to probabilistic and tailor their approaches accordingly can navigate the transformation successfully.
Success will require deliberation in choosing where to focus, realistic expectations about readiness and timelines, and significant investment in the human and organizational dimensions that enable AI to deliver its promise. The question isn’t whether AI will transform service operations, but rather which organizations will master the uneven rhythms of adoption to emerge as leaders in the AI-powered services economy.
