New Esker Research Finds Finance Leaders Want More Proof and Control Before Giving AI Greater Decision-Making Authority

Esker’s survey of 338 global finance leaders found that 72% exceeded planned AI spending over the past year, increasing pressure to demonstrate measurable ROI and strengthen governance. AI’s most cited benefit was productivity (58%), while 48% cited data quality and 42% cited finance-system integration as barriers to broader deployment; 66% of respondents also knew or suspected use of unapproved AI tools. CFOs remain particularly restrictive on autonomous AI decisions involving revenue targets and hiring budgets, signaling continued human oversight despite expanding adoption.

Analysis

This is a low-conviction sector signal rather than a company-specific catalyst: finance-software buyers appear to be moving from experimentation toward procurement discipline. Over the next 1-3 quarters, vendors whose AI monetization depends on broad seat-price increases or loosely defined “copilot” attach rates face longer sales cycles, more pilots that fail to convert, and greater pressure to document payback. That is incrementally unfavorable for premium-multiple application software with elevated AI expectations, while established workflow vendors can defend pricing only where implementation measurably reduces DSO, invoice-processing cost, fraud loss, or working-capital needs.

The non-obvious beneficiary is the control layer, not necessarily the model layer. Platforms embedded in systems of record and identity/data governance—SAP, ORCL, NOW, MSFT, plus cybersecurity and data-quality vendors such as PANW, CRWD, MSTR?—should capture incremental integration and audit spend before enterprises allow autonomous financial actions. The practical constraint is that finance data remains fragmented across ERP, procurement, treasury, CRM, and payment systems; this raises services and integration demand, favoring ACN, GLOB, and potentially MuleSoft/Salesforce ecosystem spend, but also delays recurring software revenue recognition.

Consensus remains focused on AI capex and token consumption. The more relevant 6-18 month risk is a software-budget rotation: CFOs may fund governance, data remediation, and legacy-system integration by reducing discretionary point-solution subscriptions. The survey is vendor-sponsored and small, so it does not establish a spending inflection; validate through upcoming commentary on AI pilot-to-production conversion, net retention, implementation backlog, and finance-transformation bookings. No immediate directional trade is warranted absent corroboration from large enterprise-software earnings.

Key Decisions for Investors

• Maintain a 1-3 month watchlist rather than initiate a broad AI-software short: flag NOW, CRM, WDAY, DDOG and SNOW if management reports rising AI pilot activity without corresponding paid production deployments, ARR contribution, or RPO acceleration. A guidance cut tied to elongated CFO-led approvals would be the actionable trigger.
• Prefer SAP and ORCL versus higher-multiple horizontal SaaS on a 6-12 month basis: their ERP installed bases create a tollgate for finance-data integration and governance. Structure as long SAP/ORCL versus short an equal-dollar basket of CRM and WDAY only after relative-strength confirmation; exit if cloud backlog/RPO growth decelerates materially or AI attach rates become demonstrably faster at the shorts.
• Monitor ACN and GLOB for a 2-4 quarter services-demand read-through. Initiate only if bookings/backlog show finance modernization and data-governance growth without material utilization pressure; the risk is that clients defer full transformations while limiting spend to low-margin pilots.
• Treat cybersecurity/data-governance exposure as a secondary beneficiary, not a direct AI-autonomy play. PANW and CRWD benefit only if unapproved-tool remediation converts into funded security programs; require evidence in billings, remaining performance obligations, or large-enterprise deal commentary before adding exposure.