A study covering more than 215,000 United States companies has shown that, instead of reducing the use of labour, the number of firms currently investing in artificial intelligence has continued to expand. This result contrasts with the concern that Generating AI has begun to replace white collar jobs on a large scale.
The study was completed by Ramp, the corporate finance platform, and Revelio Labs, the labour force data agency, from 2021 to early 2026. The study matched the actual records of payments made by the enterprise to the AI supplier to the employee data at the company level to observe changes in the use of AI.
High-input enterprise recruitment continues to grow
The study showed that AI enterprises with the most expenditure intensity had an average increase of about 10 per cent in the total number of employees and 12 per cent in recruitment to junior positions following the introduction of AI. In contrast, low-intensity AI-based enterprises have not experienced significant employment growth.
Ramp defines “Adoption of AI” as the payment of at least $100 per month to AI suppliers for three consecutive months. The study will then measure the intensity of adoption by the average AI expenditure per person for the first three months of the enterprise ' s deployment.
Growth is more than just technical jobs.
According to the study, new recruitment is not limited to engineering posts, but also in the areas of sales, administration, finance and customer service. The changes usually occur 6 to 12 months after an enterprise deploys AI, indicating that it takes some time for an enterprise to integrate the tools into its daily processes to see efficiency gains.
However, the research team also stressed that the results were only indicative of relevance and did not directly demonstrate that AI input itself had led to an increase in recruitment. The reason for this is that earlier AI adoption companies are already larger, more rapidly growing, more technologically oriented and more likely to receive wind investment funds.
Use still concentrated in knowledge-intensive industries
In order to reduce sample deviations, the study did not compare AI adoptions directly with never-used companies, but rather compared early adopters with enterprises that had not yet adopted but had similar characteristics.
In terms of industry distribution, the information industry has the highest AI adoption rate, followed by financial and professional services. The use of hotels, the arts and the medical industry is significantly slower, indicating that AI is still concentrated in knowledge-intensive areas.
Ramp indicated that the study was characterized by the use of real procurement data from enterprises rather than questionnaires or occupational exposure models. According to its conclusions, at this stage AI is more like helping some enterprises to expand their operations than immediately replacing large numbers of employees.