Artificial intelligence is no longer just an experiment for major companies – it’s a proven driver of business value. That’s the key finding from a new study by the Human-AI Research Department at the Wharton School, which surveyed more than 800 senior executives from large US firms. Around 75 per cent of respondents said their AI projects are already delivering a positive return on investment – a striking result amid ongoing scepticism around the real-world impact of AI.
The survey, conducted in June and July 2025, focused on companies with at least 1,000 employees and annual revenues exceeding $50 million. It aimed to capture how generative AI is being implemented in practice, with a particular emphasis on ROI, integration into core operations, and long-term strategic planning.
According to the findings, three out of four businesses reported clear financial benefits from AI initiatives, with fewer than five per cent citing negative returns. Importantly, the study suggests a shift from pilot programmes to full-scale deployment. AI is no longer confined to isolated trials – it is becoming embedded in areas like IT, procurement, marketing, and sales.
Adoption is also rising at the individual level. Eighty-two per cent of executives say they now use generative AI tools at least weekly, and 46 per cent use them daily – primarily for data analysis, content creation, and internal workflows. Strategically, AI investment is set to grow further: 88 per cent of companies plan to increase their generative AI budgets next year, with 62 per cent expecting to raise them by over 10 per cent.
Significantly, around one third of AI budgets are being allocated to internal research and development. Many firms are building customised solutions or adapting open-source tools to their own environments, with growing emphasis on security, control, and long-term scalability.
The Wharton report stands in contrast to earlier, often-cited studies – such as one from MIT claiming that 95 per cent of AI pilots fail. Independent analysts note that Wharton applies clearer criteria and focuses on mature, operationalised projects, rather than early-stage tests. That methodological difference helps explain the more optimistic outlook.
The study also highlights that ROI depends less on technical model quality than on organisational factors: leadership, skills development, and the ability to manage change. Companies that invest in training and transformation tend to see faster returns. Data protection, cybersecurity, and internal mobility are now key to sustainable success.
Cultural resistance and the challenge of scaling up successful pilots remain major hurdles. But overall, the data paints a clear picture: generative AI is gaining strategic traction and being used productively across the corporate landscape in the United States.
With global AI spending expected to reach nearly $1.5 trillion in 2025, the findings carry economic weight well beyond individual firms. The Wharton study offers a counterpoint to doom-laden headlines – and provides a grounded, evidence-based snapshot of AI’s maturing role in business.
