AI investment surges, but value and oversight lag behind
The gist
AI investments are skyrocketing, but most organizations are still flying blind when it comes to oversight, measurement, and real business value.
What to know
- By late summer 2026, 79.1% of organizations increased AI spending and 60% were piloting projects, but only 23.6% had clear AI policies in place.
- Finance teams jumped from 58% to 76% AI usage in forecasting in a year, yet just 35% can measure ROI effectively—and only 14% have a defined strategy.
- A staggering 53% of automated enterprise work now runs on untracked AI apps, leaving finance, IT, and security teams unable to connect AI spend to results.
Governance Lags AI Boom
AI spending and experimentation are outpacing policy, with most organizations lacking clear oversight even as confidence in value remains stagnant.
By early summer 2026, the pattern was already visible: AI spending and experimentation were rising faster than the systems needed to govern them. Galorath’s 2026 State of the Industry Report, released June 2, found 79.1% of organizations were increasing AI investment and 60% were actively experimenting with or piloting AI in estimation, yet only 23.6% had clear, documented AI policies. It also showed 77.7% faced data governance restrictions, 70.9% said shadow AI was common, and 51% of organizations actively increasing AI investment reported significant improvement in planning accuracy and estimation confidence, versus 11.1% of more cautious peers.
By August, Protiviti’s 2026 Global Finance Trends Survey sharpened the same contradiction inside finance: AI use for financial forecasting climbed from 58% to 76% in a year, but only 35% of finance organizations said they were effective at measuring AI ROI, and only 14% were deploying AI against a defined strategy. That aligned with KPMG’s July 2026 findings on value struggles and with the broader 2025 baseline problem: confidence had not meaningfully improved despite heavier spending and broader use, leaving governance gaps and poor ROI measurement unresolved.
Untracked AI, Unchecked Risk
Rapid AI adoption is fueling shadow operations and security gaps, leaving business impact unclear and critical governance teams sidelined.
The scaling problem starts with speed outrunning control. One July analysis captured the imbalance starkly: “86% of C-suite executives increasing their budgets” while “only 32% report a sustained impact from these investments,” and the same report found “53% of automated work in enterprise environments now runs on AI applications that aren't tracked in official systems,” meaning “unsanctioned AI activity is running below the radar of finance and IT” and outside the systems that govern spend, compliance and accountability.
As AI moves from experiments into daily operations, governance gaps become operating risks and measurement failures. KPMG said 22% of organisations had reached the “driving adoption” stage, embedding AI into workflows, but warned that “adoption without measurement” leaves gaps in value, cost control and security; that warning is reinforced by the finding that “65% of organizations currently have no dedicated defenses against this attack class” of prompt injection, showing security controls have not kept pace with deployment.
The result is fragmented rollout with no clean line to business performance. In BCG’s survey of large-company CEOs, more than half said linking AI initiatives to the P&L was a key barrier, yet only 14% had clearly defined P&L impact for all initiatives; 55% cited people redesign as a barrier but only 30% include HR in AI governance versus 82% for technology, and “nearly two-thirds” pursue pilots while only 26% have embedded AI in broader transformation, echoing Basware’s finding that 76% plan to increase investment but only 39% are ready to scale it.