Who Pays for Agentic AI? The Enterprise Budget Problem No Vendor Will Address

Why It Matters

The unresolved funding and governance model threatens costly mis‑investments and reshapes how AI value is captured across the enterprise software market.

Key Takeaways

• Transactional systems retain audit and compliance control
• Data platforms funded from analytics, not operational budgets
• Departmental owners prioritize direct productivity gains
• Vendor orchestration layers compete for AI dominance
• Governance requirements limit unbundled AI deployments

Pulse Analysis

The hype around agentic AI promises a seamless interface where natural‑language agents pull context from data lakes and execute tasks across legacy applications. In practice, the architecture consists of three layers: the execution tier (Salesforce, Workday, SAP), a unified data tier (Snowflake, Databricks), and a top‑level intelligence tier of conversational agents. While the concept is compelling, most enterprises still rely on the familiar transactional systems for day‑to‑day work, and the AI layer remains largely experimental.

Budget realities expose a structural mismatch. Data teams report to CDOs and are evaluated on insight delivery, not on funding cross‑departmental workflow automation. Conversely, business units that own CRM or ERP systems control their own spend, justifying investments that directly boost user productivity. This split leaves the orchestration layer—where AI agents would live—without a clear payer, forcing CIOs to juggle competing priorities and often resulting in fragmented pilots rather than enterprise‑wide rollouts.

For vendors, the stakes are high. Companies like Microsoft, Salesforce, and ServiceNow each leverage existing user relationships to position their own agent platforms, while vertical specialists such as SAP and Workday aim to keep AI within their domain ecosystems. Yet stringent governance—audit trails, role‑based access, and regulatory compliance—means agents must operate through, not around, the system of record. Until data platforms can match those controls, enterprises will likely adopt a hybrid model: agents augmenting transactional systems where trust and compliance are paramount, while isolated data‑cloud agents serve niche use cases. This pragmatic coexistence tempers the “great unbundling” narrative and underscores the need for clear ownership and funding strategies.