AI Can Automate 95% of Procurement Tasks, So Why Are Margins Still Under Pressure?

A profound disconnect exists at the heart of modern procurement. We are told, with figures ranging from 50% to an astonishing 95%, that Artificial Intelligence (AI) can automate the vast majority of procurement work. Yet, simultaneously, organizations are struggling with persistent margin pressure and a chronic inability to realize the promised financial returns from their technology investments. 

This is Procurement’s Silent Inefficiency Problem. It’s a systemic gap that doesn’t stem from the technology itself, but from fundamental, non-technological barriers within the enterprise. Our newest research confirms that the central inefficiency is largely a failure to address foundational prerequisites necessary for scaling AI. Simply put: you cannot automate a broken process and expect it to work better. 

The Four Pillars of Internal Resistance 

The data is overwhelmingly clear: the barrier to massive AI scaling is organizational and strategic, not technical. This “silent inefficiency” manifests in four critical, interconnected deficits that are keeping procurement stuck in a pilot purgatory. 

1. Process Complexity and Data Quality

This is the most fundamental obstacle to true transformation. A staggering 91% of executives cite entrenched process complexity as the chief barrier to scaling their AI initiatives (McNab & Mohammed, 2026 Enterprise Gen AI Key Issues Study Insights, Nov 2025). The industry has a long history of poor data governance and legacy processes. 

AI systems are being introduced into messy environments where data is often inconsistent, incomplete, and non-standardized. Automating a broken, complex process only amplifies the underlying inefficiencies. The prerequisite for success is not a bigger, more complex AI tool, but a fundamental simplification of business processes before scaling AI investments (McNab & Mohammed, 2026 Enterprise Gen AI Key Issues Study Insights, Nov 2025). 

1. Strategic Misalignment and Flawed Objectives

A significant number of AI projects fail to deliver on their expected Return on Investment (ROI) due to flawed strategy and scope creep (Tandler & Deepika, Procurement’s AI Success Starts With Setting Smarter Objectives, Oct 27, 2025). Many procurement leaders lack specific, measurable objectives, instead relying on vague goals like “efficiency” (Tandler & Deepika, Procurement’s AI Success Starts With Setting Smarter Objectives, Oct 27, 2025). 

When organizations attempt to prioritize AI objectives across all three categories—operational, tactical, and strategic—simultaneously, the risk of disappointment dramatically increases (Tandler & Deepika, Procurement’s AI Success Starts With Setting Smarter Objectives, Oct 27, 2025). This lack of focus dilutes resources and attention, confirming that managerial inefficiency is a major problem. The successful path involves starting with a narrow, clear objective and demonstrating value before attempting to scale broadly. 

1. The Talent and Literacy Gap (The Human Factor)

While AI can automate a large chunk of routine tasks, the remaining high-value work—managing the AI, handling complex exceptions, and strategic negotiation—requires specialized, hybrid skills. Low AI literacy and organizational resistance to change act as a significant brake on adoption (Ryan, Sommer & Scheibenreif, How AI-Enabled Machine Buyers Will Transform Procurement, Oct 6, 2025). 

We have long known that the “people” component accounts for approximately 70% of digital transformation success. The need for new hybrid roles focused on data science, prompt engineering, and compliance is high, but the upskilling initiatives needed for the existing workforce are often insufficient. The human element—Al system supervisors and exception monitors—is absolutely critical to mitigate risk and ensure reliability (Ryan, Sommer & Scheibenreif, How AI-Enabled Machine Buyers Will Transform Procurement, Oct 6, 2025). 

1. Governance and Risk Deficit

The shift toward autonomous agents—the idea of “machine buyers” that can execute transactions independently—introduces new, acute risks. These risks relate to security, ethics, accuracy, and trustworthiness. Critically, current governance models are not equipped to handle these new risks. 

To scale responsibly, rigorous human-in-the-loop oversight is non-negotiable. The time and investment required to establish this essential governance infrastructure—including clear governance policies and ethical considerations—detracts from immediate margin benefits but must be prioritized to ensure trust and reliability.  

Technology’s Misalignment: When a Hammer Looks for a Nail 

Another facet of the problem is technology misalignment. Many organizations risk over-investing in complex AI systems driven by hype, while neglecting proven, cost-effective automation technologies (Paradarami & Joshi, RPA’s Relevance in the AI Era of Procurement Automation, Nov 10, 2025). 

Robotic Process Automation (RPA), for example, is ideal for routine, rule-based, high-volume transactional processes (like in P2P), and in many cases, it suffices (Paradarami & Joshi, RPA’s Relevance in the AI Era of Procurement Automation, Nov 10, 2025). Deploying complex AI where a simpler RPA solution would work creates unnecessary expenditure and complexity (Paradarami & Joshi, RPA’s Relevance in the AI Era of Procurement Automation, Nov 10, 2025). A balanced strategy is necessary to match the right level of automation to the process complexity. 

The Transformation Mandate 

The organizations that are succeeding are not simply adopting AI tools; they are fundamentally reimagining their work. Success stories, like those from Johnson & Johnson and Bosch, involve using AI to redesign and simplify existing processes. 

Bosch, for instance, achieved a massive 78% efficiency increase in their P2P agentic AI workflow by focusing on process redesign, not just technology adoption (Clinton, GenAI Breakthrough: Procurement Transformation Recap Part 1, Oct 15, 2025).  

The fundamental takeaway is that AI implementation is a transformation project, not merely an IT project (Clinton, GenAI Breakthrough: Procurement Transformation Recap Part 1, Oct 15, 2025). It requires leadership commitment and organizational alignment.  

We must internalize the mandate that technological adoption alone is insufficient for success, and commit to the non-technical components—namely, process, strategy, and people—to capture margin relief. The goal is continuous change.