Memorandum
- From
- Connor Quincy via Fortune | FORTUNE
- Date
- Filed
- Business·4 min to read
- Re
Only 10% of Executives Are AI Leaders, Survey Finds
ReOnly 10% of Executives Are AI Leaders, Survey Finds
A 2026 EXL study of 322 senior decision makers found that 76% believe their companies lead in AI, while only 10% qualify as AI leaders, a gap that analysts say separates firms rebuilding their operations from those merely adding AI tools.
Most corporate executives believe their companies are ahead on artificial intelligence, but the data suggests the vast majority are not. A 2026 EXL Enterprise AI Study of 322 senior decision makers found that 76 percent consider their organizations AI leaders, while only 10 percent actually meet the criteria for that group. The 66-point gap between perception and reality is, according to the study's framing and the commentary of Collective[i] cofounder Stephen Messer, the central problem facing American businesses as AI spending accelerates.
The AI leaders in the study showed measurable results. On average, they reported 27 percent revenue growth, 26 percent cost savings and a 22 percent improvement in margins. Those figures distinguish companies that have rebuilt their operating models around AI from those that have layered new software onto existing processes. Messer argues that the difference is less about which model a company buys and more about how it restructures work around it.
That distinction has been seen before in research on enterprise AI adoption. An August 2025 report from the MIT Media Lab's NANDA initiative, led by Aditya Challapally, examined 150 enterprise leaders, 350 employees and 300 public AI deployments. It found that 95 percent of pilot programs stalled before reaching scale. Challapally described the cause as a «l«earning gap,» with winners redesigning workflows around the tool and losers fitting the tool into workflows that never changed.
Messer frames the divide as a “K-shaped” economy, a term economist Peter Atwater coined in 2020 to describe the pandemic recovery that split one economy into two diverging paths. In his telling, one group of companies is rebuilding what he calls the operating substrate, a stack of different AI systems each aimed at a specific problem and orchestrated together. The other group is purchasing AI seats or chatbot features and treating the purchase as a strategy. Messer says buying a single language model subscription is the easy part; running a substrate is the real work.
The stakes are being shaped by the financial markets as well. Anthropic filed its public S-1 on October 26, 2026, to list on Nasdaq under the ticker ANT, with the roadshow starting November 9 and the debut expected before Thanksgiving. The company is reportedly targeting a valuation of about $1.5 trillion, down from a $2 trillion figure floated in October. Anthropic has reported a roughly $68 billion annualized revenue rate, with about 1,500 employees, and roughly $331 billion in stock awards on its balance sheet. Its nine-month revenue reached $33.2 billion at an 85.5 percent gross margin, and the number of customers paying $100,000 or more a year rose from about 950 to more than 11,900 in twelve months.
The offering would be the third major AI exit of the second half of 2026, following SpaceX's $60 billion acquisition of Cursor in August and AMD's $8.2 billion purchase of World Labs, founded by Fei-Fei Li, in September. OpenAI, valued at about $1.25 trillion with roughly 70 percent quarterly annualized growth, sits in the same market. Messer's argument is that these valuations reflect investor confidence that AI will reshape the operating foundations of most major companies over the next decade, not simply that one model will win.
For executives, the findings raise a practical question about whether their companies are part of the minority that is changing how work gets done or part of the majority that has purchased the tools without changing the work. The EXL data suggests that the confidence most leaders feel about their AI position is not yet matched by results. Whether the gap narrows will depend on how many organizations move from buying AI to rebuilding around it.
46
Source: Collective[i] cofounder: 76% of executives think their companies are AI leaders. Only 10% are
