The most-quoted AI agent statistics count intent, and real use in any single department is no more than 10 percent, finds a source-checked analysis by bdautomated
75 figures from 18 publishers traced to the original reports, with who was asked and what counted; the dataset is free to download
15 September 2026. bdautomated today published “AI agent statistics 2026: every number checked at its source”, a free reference page and dataset that traces 75 widely quoted statistics about AI agents and AI use in business back to the documents they came from.
The analysis finds that the numbers disagree because they count different things. In McKinsey’s 2025 global survey, 62 percent of organizations were at least experimenting with AI agents, 23 percent had scaled one somewhere in the company, and in any single business function no more than 10 percent had. Surveys that count any adoption or intent report far higher figures: 79 percent of U.S. executives told PwC in April 2025 that agents were already being adopted in their companies, while a Capgemini survey that re-checked what respondents meant by “agent” found 14 percent had implemented one. Across all U.S. businesses of every size, the Census Bureau found 19.8 percent using AI in any business function as of May 2026.
The page also decodes the figure that rattled markets in 2025. MIT Project NANDA’s finding that “95 percent of organizations are getting zero return” measured profit-and-loss impact within roughly six months of a pilot, in a sample of 52 interviews, 153 conference survey responses and 300 public deployments, and its authors call the findings preliminary. It does not say that 95 percent of AI projects fail. Gartner’s prediction that over 40 percent of agentic AI projects will be canceled by the end of 2027 is a forecast from June 2025; nobody has counted the cancellations yet.
Every figure on the page passed four checks: the number appears in the original document; the exact place and a verbatim quote are recorded; what it measures is written in plain words, including who was asked, how many and when; and it is set against the other sources, with disagreements shown rather than averaged. Market-size forecasts were left out because the reports behind them are paid and cannot be checked. The dataset is published as CSV and JSON under a CC BY 4.0 licence, and the page carries a corrections address.
“Two headlines in the same week said almost nobody has AI agents running and almost everybody does, and both were quoting real surveys. We wanted the page we could not find: what each survey actually asked, so a business owner can tell which number is about a company like theirs,” said a spokesperson for bdautomated.
The page includes four charts that other publications may embed with attribution, and a table of all 75 figures with their sources, dates, samples and quotes.
Read the analysis: https://bdautomated.com/ai-agent-statistics/
checked 2026-09-15 webfetch: “AI Agent Statistics 2026: Every Number Checked at Its Source”
Download the data: https://bdautomated.com/data/ai-agent-statistics.csv
checked 2026-09-15 webfetch: “id,cat,value,num,stat,counts,src,doc,date,sample,place,quote,url,verdict”
About bdautomated
bdautomated builds The Book, a named staff of AI agents installed in the accounts a company already runs on. The agents research, prepare, draft, chase and check, each at the level of autonomy the customer sets. bdautomated hosts nothing and stores no customer records.
https://bdautomated.com/
Media contact
