Facing tight timelines to deploy AI, drive its internal adoption and ensure governance, over half of senior IT and Finance leaders question whether their AI investments are delivering expected results.

Larridin, the leading AI measurement and scaling platform, today released its Larridin 2025 State of Enterprise AI Report, a July study of 350 senior Finance and IT leaders at companies with 1,000+ employees, revealing how enterprises are using AI, how it is affecting business outcomes, and the growing risks of ungoverned AI usage.

Download the full survey report here

The AI Adoption Challenge

With AI investments projected to reach $644 billion by year end 2025, enterprises face a critical hurdle: driving adoption. The core issue is assessing, tracking, and managing how AI is actually used, raising questions about whether organizations are harnessing the full value of their investments in this tech. The Report findings are clear: 72% of leaders believe AI is impacting profitability. Rather than delivering performance gains, the lack of scale and optimization of applications is putting companies at risk of falling behind the competition.

In addition, companies are racing against the clock, with 85% believing they have just 18 months to build AI capabilities before falling behind. This urgency widens the gap between ambition and accountability, driving investments in costly solutions that often outpace employees’ readiness to adapt, preventing organizations from realizing the full benefits of AI adoption.

The survey highlights enterprise leaders’ perspectives on the AI vs. ROI challenge:

• 55% are unsure if their investments are paying off.
• Nearly two-thirds admit to having lost visibility into their AI stack.
• 78% want AI transformation but fear losing control over implementation.
• Eight in ten worry that confidential data is being exposed to public AI models.
• Almost 80% report that untracked AI budgets are a growing accounting concern.

These issues are even more pronounced in large enterprises (10,000+ employees), compromising organizational agility, employee performance, and financial health.

“AI is growing at an incredible pace; however, companies lose sight of it even faster,” said Russ Fradin, CEO and Co-Founder of Larridin. “To unlock AI’s full potential and accelerate internal adoption, organizations must gain visibility into their stack by doing a full AI diagnosis that entails implementing systems to track, measure, and optimize usage. This will allow enterprises to increase ROI, boost productivity, and scale the best use cases, improving their overall health.”

The Paradox of Security Risks

Larridin’s report also reveals a paradox: while 91% of enterprises conduct internal AI audits regularly, leaders are still losing visibility into the tools being used. Over 84% discovered more AI applications than expected, signaling that employees are bypassing official channels to adopt solutions they believe are needed without IT or executive oversight.

The lack of visibility brings another set of risks driven by employees independently adopting and installing AI applications without authorization. These risks further complicate ROI measurement, an issue occurring in 81% of enterprises and cited as the top governance challenge by half of respondents.

The survey also identified the most common ways employees bypass AI governance controls:

• Using browser-based AI tools that aren't tracked (48%).
• Adopting free or public AI models (45%).
• Using personal accounts to access AI tools (44%).

This situation highlights a growing internal governance disconnect: security capabilities are failing to evolve as quickly as the adoption of unsanctioned AI. The top three security concerns cited in the study are:

1. 52% - Accidental sharing of sensitive data.
2. 44% - Outdated or insufficient security protocols.
3. 44% - Shadow AI tools installed without approval.

“Employees want to leverage AI, but the steep learning curve often pushes them toward tools they already know,” added Fradin. “Solving the security and broader challenges of AI adoption demands more than just the right AI stack and infrastructure. It also requires a strategy grounded in evidence-based measurement and validation that can enforce internal policies and allocate investments where they are actually needed to improve overall performance.”

About the Larridin 2025 State of Enterprise AI Study Methodology

The research study was conducted by TrendCandy on behalf of Larridin in July 2025, using a random sample of 350 senior Finance and IT leaders in companies with 1,000 or more employees. Download the full survey report here

About Larridin

With $17M in seed funding led by Andreessen Horowitz, Bloomberg Beta, Google Ventures, Larridin fulfills the promise of Enterprise AI with measurement, enablement and governance solutions that help companies optimize adoption, enforce internal policy, and drive strategic AI investments. Larridin provides the complete AI execution intelligence system that transforms enterprise AI from chaos to competitive advantage—revealing what's happening, connecting everything strategically, and proving measurable execution impact. Scout reveals your AI landscape, Nexus scales everything strategically, and Vantage proves measurable impact on business execution. Follow us on LinkedIn and X.

View source version on businesswire.com: https://www.businesswire.com/news/home/20250916598732/en/