Three years ago, generative AI tools like ChatGPT started a new era of artificial intelligence (AI), and the technology has since become increasingly commonplace in organizations and the general public. In 2025, we entered the era of autonomous AI agents, where major corporations have started adopting real agentic workflows. However, a new report from McKinsey, "The State of AI in 2025: Agents, Innovation, and Transformation," shows a significant gap between the widespread adoption of AI and its actual impact on enterprise-level value.
While nearly nine out of ten organizations are using AI in their workflow, most are still in the early stages of scaling the technology and realizing its full potential. The report, based on a survey of 1,993 participants across 105 nations, provides a comprehensive overview of the current state of AI adoption, highlighting key trends, challenges, and opportunities for businesses. The report analyzes the rise of AI agents, the impact of AI on innovation and business transformation, and the growing perspectives on AI's effect on the workforce.
Motion: Create hundreds of AI employees and manage all human and AI work in a single productivity-rich work management platform.
McKinsey's The State of AI in 2025: Agents, Innovation, and Transformation Key Findings:
AI Adoption is Widespread, but Scaling Remains a Challenge:
Nearly nine in ten organizations (88%) now use AI in at least one business function, which is up from 78% a year ago. Yet only about a third say they've started to scale AI across the enterprise. Put differently, adoption is broad, the impact is still uneven, and the hard work is between proof of concept and process redesign. High performers succeed because they treat AI as a tool for growth and new innovation, not just for increasing efficiency, and they're explicitly reorganizing their workflows to capture that upside.
High Curiosity in AI Agents:
2025 was the year of AI agents. As most people already know by now, AI agents are autonomous systems that can plan and take multi-step actions. According to McKinsey's findings, 62% of organizations are at least experimenting with agents (23% scaling in one or two functions; 39% experimenting). That's a big jump in seriousness, even if scaling is still concentrated in roles like IT service desks and knowledge management, and adoption is most common in tech, media/telecom, and healthcare.
Positive Leading Indicators on AI's Impact:
The report has also shown some positive signs of AI's impact on a use-case level, with respondents reporting cost benefits in software engineering, manufacturing, & IT, and revenue benefits in marketing/sales, strategy, & product development. Additionally, 64% of respondents say that AI is allowing innovation within their organizations.
However, only 39% report a real impact on EBIT at the enterprise level, indicating that the full financial benefits of AI are yet to be realized.
High Performers Use AI for Growth and Innovation:
The companies that are seeing the most value from AI are those that use it to drive growth and innovation, in addition to efficiency. These "AI high performers" are more likely to set transformative goals for their AI initiatives and redesign their workflows to take full advantage of the technology's potential.
Differing Perspectives on Employment Impact:
The report highlights the inconsistent expectations of AI's impact on the workforce. While 43% of respondents expect no change in the overall size of their organization's workforce in the coming year, 32% expect decreases, and 13% expect increases. This suggests that the impact of AI on employment is likely to be complicated and multifaceted, with some roles being displaced while new ones are created.
Otter.ai: An AI-powered meeting assistant that transcribes conversations in real time, generates automated summaries, and action items.
What's actually changing inside companies
McKinsey's data points to a subtle organizational rewiring with more cross-functional operating models, workflow redesign (not just tool swaps), and senior leadership attention to AI governance.
- Larger companies are leading the way—almost half of firms with over $5 billion in revenue say they are in the scaling phase. The top-performing companies invest more than 20% of their digital budgets in AI to keep growing.
In short: Strategy, structure, and spend are coming together.
- Risk management is growing up, too. A slim majority of organizations using AI report having experienced at least one negative consequence, most commonly inaccuracy; mitigation efforts are growing (privacy, explainability, reputation, compliance), but explainability is still lagging.
The pragmatic read: If you want to scale agents this year, your model-risk playbook and evaluation pipelines need to be as real as your product roadmap.
The Path Forward: From Experimentation to Transformation
The McKinsey report makes it clear that the journey to AI-powered transformation is a marathon, not a sprint. While our initial excitement around generative AI has led to widespread experimentation, the real challenge is in scaling the technology and integrating it into business workflows. To do so, organizations should adopt a more ambitious vision for AI-driven growth and innovation rather than focusing purely on an efficiency-focused mindset.
This requires a multi-pronged approach that includes:
- Redesigning workflows: To fully take advantage of the power of AI, organizations need to fundamentally rethink their processes and workflows. This involves identifying opportunities to automate tasks, augment human capabilities, and create new forms of value.
- Investing in talent: The demand for AI-related skills is on the rise, and organizations need to invest in up-skilling and re-skilling their workforce to meet this demand. This includes hiring for roles such as machine learning engineers, data engineers, and data scientists, as well as providing training to existing employees.
- Encouraging a culture of innovation: To succeed with AI, organizations need to create a culture that encourages experimentation, learning, and collaboration. This involves breaking down silos, allowing employees to take risks, and celebrating both successes and failures.
If 2023 was the breakout and 2024 the adoption spike, 2025 is the operational turn. Organizations report using AI in multiple functions, and the "where" is telling; alongside traditional areas like IT and marketing/sales, knowledge management is becoming a focus. You can feel the center of gravity moving from "ask a chatbot" to "let the system capture, route, and act on knowledge." That's where agents naturally fit in, automating multi-step, cross-tool work that used to need a person in the loop.
In Conclusion:
The state of AI in 2025 isn't a story about whether companies use AI—they do; it is a story of AI's progress and potential, and whether AI is in the flow of work. McKinsey's latest report showed that AI agents moving from experiment to execution, value is found in well-designed use cases, and different leadership patterns separate those who just experiment from those who truly transform. If you're serious this year, don't ask "Which AI model?" first. Ask "Which process gets redesigned, by whom, with what guardrails?" Answer that, and the AI agents, budgets, and benefits will follow.
💡 For Partnership/Promotion on AI Tools Club, please check out our partnership page.
