A 2026 survey by Writer, conducted with research firm Workplace Intelligence among 2,400 executives and employees at large global companies, points to a clear gap: 97% of executives say their company has benefited from AI, yet only 29% report a significant financial return from generative AI. Heavy AI users, according to the executives surveyed, are at least five times more productive than their peers. Sure, Writer sells enterprise AI products, so it has a direct commercial stake in how these numbers get read. The gap between individual productivity and company-level financial results shows up, though, at similar levels, in independent research from the same period.
Heavy AI users, according to the executives surveyed, are at least five times more productive than their peers. Writer sells enterprise AI products, a direct commercial stake in the outcome. The adoption-to-ROI gap it reports nonetheless tracks with independent research published in the same period.
Seventy-nine percent of organizations surveyed report difficulty adopting AI, a double-digit jump from last year. Nearly half (48%) call AI adoption a "massive disappointment" — up from 34% in 2025. Three-quarters of executives admit their AI strategy is more for show than an actual internal guide to action.
Old workflows cap the gains
A separate survey, run in April 2026 by Futurum Group among 830 IT decision-makers, shows a shift in expectations: companies have stopped accepting "productivity" alone as sufficient justification for AI spending. The share of companies demanding direct financial impact — revenue growth or margin improvement — nearly doubled to 21.7% of responses, while productivity fell as the top metric, from 23.8% to 18%. The market wants proof, not just activity.
The gap between individual output and company-level results is explained directly in a CIO.com piece by enterprise architect Jeff Carson: employees get faster individually, then get sent back into the same meetings, approvals, and reporting layers that slowed work down before AI arrived. The tool is new. The process around it isn't.
Training has to match the role
A marketing employee, a software engineer, and an HR manager face different risks, different data sources, and different approval processes. A generic "introduction to AI" course addresses none of that specifically.
Deloitte's latest State of AI in the Enterprise report, based on responses from 3,235 leaders across 24 countries, identifies insufficient worker skills as the most commonly cited barrier to integrating AI into existing workflows. Two-thirds of companies (66%) already report productivity and efficiency gains. Revenue growth, by contrast, remains mostly aspirational: 74% of companies are pursuing it, but only 20% have actually achieved it.
Data governance and unclear policy
A concrete finding confirms the risk: 67% of the executives surveyed by Writer believe their company has already suffered a data leak or security breach caused by an unapproved AI tool. 35% of employees have entered proprietary company information into public AI tools. More than a third of companies (36%) have no formal plan for supervising AI agents, and 35% admit they couldn't immediately shut down a malfunctioning agent.
Managers are the least trained link
Only 35% of employees consider their manager an actual AI champion at work. 58% of executives admit that many of their fellow leaders lack the basic knowledge needed to make strategic AI decisions.
The conclusion that emerges from combining these sources has little to do with the technology's potential. That's already confirmed by the individual productivity gains nearly every surveyed company reports. It has to do with the fact that most companies still measure AI adoption in licenses purchased and agents deployed, not in decision speed or the additional revenue AI was supposed to generate.
Sources: Writer / Workplace Intelligence, "Enterprise AI Adoption in 2026"; Futurum Group, 1H 2026 Enterprise Software Decision Maker Survey; Deloitte, "State of AI in the Enterprise" (2026); CIO.com.