Scaling AI agents for real enterprise impact
The McKinsey Global Survey 2025 reveals that while 88% of organizations have adopted AI, only a fraction have successfully scaled the technology to achieve significant enterprise-wide financial impact. The emergence of AI agents marks a new frontier in automation, yet moving beyond mere experimentation requires a fundamental redesign of business workflows and stronger leadership commitment. High-performing companies differentiate themselves by viewing AI as a catalyst for transformative innovation rather than just a tool for incremental efficiency.
Points clés
- 88% of organizations report regular AI use in 2025, up from 78% in the previous year.
- 62% of organizations are currently experimenting with AI agents capable of executing multi-step workflows.
- Only 39% of respondents report a significant impact on EBIT at the enterprise level despite high adoption rates.
- Large companies with over $5 billion in revenue are nearly twice as likely to have scaled AI compared to smaller firms.
- AI high performers, representing 6% of the sample, attribute more than 5% of their EBIT to AI initiatives.
- High-performing organizations are 3.6 times more likely to seek transformative change over incremental efficiency.
- 32% of respondents anticipate an overall workforce reduction of 3% or more in the coming year due to AI integration.
- High performers allocate more than 20% of their digital budgets specifically to AI development and deployment.
- Organizations are managing an average of four distinct AI risks today, doubling the figure from 2022.
- The McKinsey survey was based on data from 1,993 global participants across diverse industries.
À retenir
So, everyone is “doing” AI now, but apparently, most are just poking it with a stick to see if it moves. If you want to join the elite 6% who actually make money from this, you might want to try actually redesigning your messy workflows instead of just sprinkling “agent” fairy dust on top of them. Or, you could just keep running pilots until the heat death of the universe; at least it keeps the IT department busy while the big players prepare to automate a third of the workforce. Good luck with those “human-in-the-loop” processes—let’s hope the humans stay in the loop long enough to see the ROI appear!
Sources
Quiz sur le document: 10 questions
