Leah Unveils New Research on Enterprise AI Success; 79% of Organizations Report AI Governance Blind Spots

IDC research commissioned by enterprise agentic-AI company Leah found that 66% of organizations already use AI agents in production and expect deployed agent counts to increase sixfold by early 2027. The report identifies governance, interoperability and cross-functional coordination as key constraints on scaling agentic AI, indicating strong adoption momentum but material execution and control challenges.

Analysis

This is not yet an investable demand datapoint: vendor-sponsored survey evidence tends to measure intent rather than incremental budget, and agent deployment counts can expand through low-value workflow pilots without materially changing software spend. The more relevant read-through is that enterprise adoption is likely to shift from standalone model access toward control-plane spending—identity, permissions, auditability, observability, data integration, and workflow orchestration—before broad autonomous-agent ROI is proven.

Near-term beneficiaries are likely incumbents that already sit in enterprise systems of record: MSFT (Entra, Azure, Copilot), NOW (workflow governance), CRM (customer-data/workflow layer), PANW and CRWD (agent security), plus Datadog (DDOG) if agent observability becomes a discrete budget line. The risk to pure-play agent vendors is that governance requirements reinforce platform consolidation: buyers may prefer bundled capabilities from hyperscalers and existing SaaS vendors rather than adding another unproven orchestration layer. Over 6-18 months, interoperability standards could reduce switching costs and pressure agent-platform pricing, while expanding the addressable market for integration vendors such as Mulesoft/CRM, ServiceNow and IBM.

Consensus remains overly focused on token consumption and model quality; the potential bottleneck is organizational liability. A high-profile unauthorized-action, data-leakage, or audit failure would likely delay autonomous deployments and redirect budgets to security and governance, benefiting PANW/CRWD but impairing application-layer AI multiples. The thesis is falsified if major enterprise software vendors report sustained AI attach rates without a corresponding increase in security, integration, or professional-services demand, indicating that deployments remain narrow and centrally managed rather than operationally scaled.

Key Decisions for Investors

• No directional trade on the survey or Leah; treat it as weak, non-independent evidence rather than a catalyst.
• Build a 1-3 month watch basket of NOW, PANW, CRWD, DDOG, MSFT and CRM; look for earnings evidence of AI-related bookings, security-module attach, or services backlog before adding exposure.
• Prefer a relative-value expression of long NOW or PANW versus a basket of high-multiple, smaller AI application vendors if evidence emerges that enterprise procurement is consolidating around governed platforms; target a 10-15% relative move over 6-12 months, with exit on weakening subscription growth or explicit AI-monetization guidance cuts.
• Monitor enterprise AI security incidents and regulatory guidance over the next two quarters. A material incident is a potential short-term catalyst for PANW/CRWD outperformance, but broad AI-software multiple compression would make outright long exposure less attractive.