As artificial intelligence reshapes corporate operations across New York, procurement teams are finding themselves on the front line of a new governance challenge — one that traditional buying processes were never designed to handle. Industry analysts and governance experts say the shift has transformed procurement from a commercial function into one of the most consequential and least mature components of enterprise AI strategy.
The core problem is an asymmetry of expertise and power. Procurement teams, typically trained to negotiate software contracts on price and service levels, now face decisions about highly complex AI systems that raise questions about data ownership, model training, liability for errors, and intellectual property exposure. Most teams lack the technical depth to audit these systems before making purchasing decisions.
Market concentration compounds the challenge. OpenAI, Anthropic, and Google collectively account for approximately 88 percent of enterprise large language model usage, according to Menlo Ventures. This concentration gives dominant providers leverage to set terms and implement changes without meaningful negotiation, leaving buyers with limited bargaining power.
The nature of AI systems also creates novel post-contract risks. Rolling updates and feature releases can alter accountability structures and introduce new risks after agreements are signed. AI supply chains span models, infrastructure providers, APIs, and application layers, making it difficult to assign liability when problems arise.
Data management is among the most underestimated areas of risk, experts warn. Core due diligence questions about data collection, storage, and model training are frequently left unanswered at the point of contract, potentially exposing organizations to the inadvertent disclosure of confidential information or customer records to external training processes.
For New York’s dense concentration of financial services, media, and professional services firms — all heavy AI adopters — the governance gap in procurement carries outsized risk. Industry groups are beginning to develop AI-specific procurement frameworks, but adoption remains uneven as organizations balance the urgency of AI deployment against the need for proper governance.
Sources: New York Today, Menlo Ventures Enterprise AI Report
