Recent reports show that while enterprise AI is indeed gaining traction within large enterprises, it is mostly limited to a narrow range of tasks that are easy to measure, automate, and validate.
According to US-based venture capital firm Andreessen Horowitz, 29% of Fortune 500 companies and about 19% of Global 2000 companies are currently in operation, paying customers of major AI startups based on deployments that involve top-down contracts, pilot conversions, and actual use within their organizations.
High momentum in restricted workflows
However, progress has not been spread evenly across the enterprise. Andreessen Horowitz said that coding, customer support, and search account for the largest share of enterprise AI use cases today, and while coding is the clear outlier, the areas with the strongest initial demand are technology, law, and healthcare.
The company said it mapped revenue momentum by use case against OpenAI’s GDPval benchmark. OpenAI says GDPval measures model performance on 1,320 specialized tasks across 44 occupations, including tasks based on real-world work products.
Andreessen-Horowitz said support has been found to be particularly receptive because tasks have deadlines, constrained objectives, and are easy to measure success through metrics such as ticket volume, customer satisfaction, and resolution rates.
More broadly, we find that the fastest-advancing fields tend to be text-based, iterative, reviewable, and easily verifiable, while work shaped by the constraints of the physical world, tighter regulations, interpersonal judgment, or coordination among multiple parties has slower adoption.
Transition to operational discipline
Another 2026 study by WRITER and Workplace Intelligence suggests that as companies move beyond pilots, the more challenging issue will be work discipline. The study surveyed 2,400 knowledge workers (including 1,200 executives and 1,200 employees) in the US, UK, Ireland, Benelux, France, and Germany.
This paper focuses on those already using or allowing generative AI in the workplace, and provides a perspective on organizations that are already past the awareness stage.
The survey found that for many respondents, adoption is already part of their daily routine. WRITER and Workplace Intelligence reported that 94% of executive respondents use AI tools for at least 30 minutes per day, and 64% spend at least 2 hours per day using generative AI tools or AI agents.
According to the same report, 52% of employees say they have used an AI agent in the past year, and 97% of executives say their company has implemented an AI agent in that time.
High utilization results in uneven revenue
But the same organizations have struggled to prove broader benefits. According to a study by WRITER and Workplace Intelligence, only 29% of executives say they have seen a significant ROI from generative AI, and 48% say implementing AI in their organization has been a “huge disappointment.”
Additionally, 39% say their organization does not yet have a formal AI strategy to drive revenue, suggesting that usage is growing faster than operating models are maturing.
Data risk concerns and organizational silos
Control gaps were also observed throughout the study. According to WRITER and Workplace Intelligence, 35% of employees admitted to entering their company’s proprietary, confidential, or confidential information into public AI tools.
At the C-suite level, 67% of executives said they believe their company has already suffered a data breach or security breach because an employee used an unapproved AI tool. The survey also found that 79% of executives say AI applications are being deployed in silos, and 54% say AI deployments will “tear apart their companies.”
Beyond the pilot stage
Taken together, the two reports demonstrate that while the enterprise market remains in the experimental phase, it is far from established. The clearest advances are occurring in a small number of workflows that can insert AI into defined tasks and make decisions based on visible results.
The broader challenge now is whether companies can build the governance, security controls, and operational discipline needed to turn early successes into lasting ones.
