Ramp 用 3 万多家企业的交易流水统计美国企业 AI 采用率
How we built The Ramp AI Index Our dataset is built by extracting text from billions of aggregated, anonymized transactions from over 30,000 businesses using Ramp Bill Pay and corporate cards
Ramp 在 2025 年春季商业支出报告中发布 Ramp AI Index,从 3 万多家使用 Ramp Bill Pay 和对公卡的企业中提取数十亿笔匿名交易文本,按商户名和账单明细识别某月是否购买过 AI 产品或服务,以此判断企业是否采用 AI。
Our Spring 2025 Business Spending Report introduces a new piece of ongoing research, The Ramp AI Index, which looks at how U.S. businesses are using artificial intelligence products and services.
AI may be the transformative technology that breaks the United States out of a two-decade slowdown in labor productivity. If AI is going to drive economic growth, its adoption will be the first leading indicator. There are datasets available that attempt to track AI adoption, but none of them use actual transaction data to do so.
Ours does.

Download our free Business Spending Report
The methodology
The Ramp AI Index builds on previous work by Bonney et al. (2024) by providing a new dataset to measure business adoption of AI. Previous work has relied almost entirely on surveys that ask businesses if they use AI. However, surveys can lead to underreporting, particularly when the questions are unclear or when the adoption of new technology is changing very quickly. Using contract and transaction data from corporate spend with AI companies, we produced a more timely and accurate measurement of AI adoption by U.S. businesses.
Our dataset is built by extracting text from billions of aggregated, anonymized transactions from over 30,000 businesses using Ramp Bill Pay and corporate cards. We consider businesses to have adopted AI if they have a transaction for an AI product or service, identified using merchant name and line-item details from receipts and bills, in a given month.
The results
In producing this first edition of The Ramp AI Index, we discovered that the U.S. government is vastly underreporting AI adoption.
The Census Bureau survey asks, “Does your business use artificial intelligence to produce goods and services?” We think this question to be broad and abstract. Does it include customer service representatives powered by AI? Sales chatbots? What about software engineers coding with AI? Or does it mean using AI in the literal manufacturing of goods? The question forces respondents to draw an arbitrary line.
Our estimate, on the other hand, uses the methodology outlined above. We find that a significantly higher percentage of corporations already use AI, even if they tell the Census Bureau they aren’t.
Critics of this research will point out that our customers are more likely to adopt AI solutions anyway since they already use Ramp, an AI-oriented approach, as their business spend platform. It’s a fair criticism. But it’s worth noting that this data may still underestimate the actual AI adoption rate among businesses, as our results don’t include usage of free AI tools or employees’ usage of personal AI accounts.
In any case, we know that businesses are adopting AI in droves—and fast.
Ramp is dedicated to helping businesses of all sizes make informed decisions. We adhere to strict
to ensure that our content meets and maintains our high standards.
来源:Ramp AI Index · ramp.com