How Organizations Are Rewiring to Capture Value from AI—Insights from McKinsey
(Note: ChatGPT assistance was used for writing this article).
Artificial intelligence (AI) continues to accelerate at a breathtaking pace, and enterprises large and small are wrestling with how to position themselves for success. McKinsey & Company’s latest report, The State of AI: How Organizations Are Rewiring to Capture Value (March 2025), provides a detailed snapshot of how businesses across industries are adopting both analytical AI and generative AI (gen AI), the hurdles they face, and the best practices that deliver tangible benefits. Below is an overview of the most important findings—along with my thoughts on why they matter for businesses seeking to harness AI’s capabilities responsibly and effectively.
1. AI Adoption Is Spreading Quickly Across Functions
One of the most striking takeaways from McKinsey’s research is the breadth and speed of AI adoption. According to the report:
These numbers reflect a significant jump in just a short time. AI usage is no longer siloed in research labs or specialized data teams; it has entered marketing, sales, IT, operations, risk management, product development, and even knowledge management. The most prevalent applications include natural-language generation, image creation, and computer code generation—though usage patterns vary by industry.
My opinion: We are past the point of “wait and see.” Even smaller firms can’t afford to stay on the sidelines while larger competitors leverage AI in many aspects of their operations. That said, adopting AI is not merely a matter of purchasing software. Rather, it requires rethinking how each function operates, from strategic planning and data governance to daily workflows.
2. Larger Enterprises Are Investing More Deeply
The McKinsey report notes that companies with annual revenues of $500 million or more are taking more decisive steps than smaller organizations. They tend to:
About 28% of surveyed organizations say their CEO personally oversees AI governance—though among larger companies, the board of directors or a senior C-suite subcommittee often shares responsibility.
My opinion: Executive buy-in and direct involvement are crucial. AI transforms everything from cost structures to cultural norms, so if it’s confined to an IT or data team, it rarely yields company-wide impact. I see this as a consistent theme: the more strategic and visible the AI leadership, the faster the organization reaps AI’s potential.
3. Workflow Redesign Is the Key to Real Impact
McKinsey highlights the importance of rewiring how work actually gets done. The report shows:
“Out of 25 attributes tested for organizations of all sizes, the redesign of workflows has the biggest effect on an organization’s ability to see EBIT impact from its use of gen AI.”
In other words, implementing AI in a superficial way—where employees simply bolt a model onto an old process—delivers only incremental improvements. The real payoff emerges when companies step back, map out how tasks flow from one department to another, and then re-engineer that flow to maximize AI’s strengths. Currently, only 21% of survey respondents using gen AI say their organizations have “fundamentally redesigned” at least some workflows, suggesting many are still in the early adoption phase.
My opinion: I see so many organizations adopt AI pilots that deliver only partial ROI because they’re not complemented by broader process changes or training. AI thrives in an environment where relevant data is available at the right time, tasks are orchestrated efficiently, and people trust and understand how to work with machine-generated outputs. In practice, this can mean reorganizing teams, changing job responsibilities, or even launching new product lines. It’s a heavier lift but absolutely necessary for meaningful returns.
4. Risk Management Is Becoming More Centralized and Proactive
As AI usage expands, so do concerns about inaccuracy, bias, cybersecurity risks, intellectual property infringement, and regulatory compliance. McKinsey’s findings reveal that many companies have centralized their AI risk and compliance functions—often in a hub or center of excellence that sets standards across the enterprise. Survey respondents are increasingly focused on mitigating inaccuracy and intellectual property risks, in particular. They are also putting humans “in the loop” to review AI outputs (though there’s wide variation in how much content gets reviewed, from 20% or less to nearly 100%).
My opinion: Risk management can’t be an afterthought. AI governance policies should be in place before major rollouts, particularly for tools that interact with customers or handle sensitive data. I believe a well-designed governance framework can actually speed up AI adoption by reassuring stakeholders (including regulators, boards, and front-line employees) that the company is prepared to handle AI’s pitfalls responsibly.
5. Best Practices for AI Adoption and Scaling
Based on McKinsey’s research, AI doesn’t drive significant value until certain best practices are in place. The report cites 12 adoption and scaling practices that correlate with higher EBIT impact. A few that stand out:
Less than one-third of respondents say their organization follows most of these practices. Even fewer—well under 20%—report tracking specific gen AI KPIs. Larger companies fare better than smaller ones at implementing these practices.
My opinion: These practices might seem like standard change-management concepts, but they’re surprisingly lacking in many AI initiatives. That signals a major opportunity for companies to surpass their peers. In my experience, the difference between a successful AI rollout and a stalled experiment often boils down to having a clear plan, leadership that drives adoption, and well-structured measurement systems.
6. AI’s Influence on the Workforce: New Roles and Reskilling
Organizations continue to hire for AI-related roles—data scientists, machine learning engineers, AI product managers—and they also need newly emerging specialties such as AI compliance and AI ethics. While it remains challenging to find skilled talent, McKinsey reports that it’s not quite as difficult as in prior years.
Meanwhile, there is a big push toward reskilling existing employees. A large share of respondents expect to retrain more workers over the next three years than they retrained over the past year. Businesses are also learning how to handle the time savings from AI—some reassign employees to new, higher-value projects, while others trim headcount in functions like service operations or supply-chain management.
My opinion: Talent is the lynchpin of any AI strategy. As certain repetitive tasks are automated, new opportunities arise for employees with deeper analytical or creative skills, or for those who can interpret AI outputs in ways that deliver business insight. This reallocation of tasks can be a delicate process; companies should provide training pathways that help employees move into higher-impact roles. Otherwise, they might face skill gaps or cultural pushback.
7. Revenue and Cost Reductions: Early Wins, Limited Enterprise-Wide Impact
Many respondents indicate that AI initiatives have already boosted revenue and cut costs within specific business units. Marketing and sales, for instance, report measurable top-line gains from personalization and content-generation tools. IT often sees cost savings from automated coding or bug detection. However, over 80% of respondents have not yet seen a major impact on overall EBIT at the enterprise level.
My opinion: Given how recent many AI deployments are, an enterprise-wide transformation can take months or years. The bright side is that any proven successes—e.g., a marketing pilot that drove a clear revenue bump—can build the momentum and confidence needed to scale AI across the organization. It’s the classic “start small, think big” approach that often underlies major technological shifts.
8. Looking Ahead: Building the “Rewired” Organization
The central theme of McKinsey’s report is that rewiring organizations—rather than layering on AI—makes all the difference. A truly AI-enabled enterprise has:
My opinion: Companies that want to stay competitive should build or reinforce these pillars now. As AI rapidly evolves—especially in the realm of generative and “agentic” AI—prepared organizations will be better positioned to capitalize on new breakthroughs. It’s likely that the gap between AI leaders and laggards will only grow as more businesses adopt these transformations in 2025 and beyond.
Conclusion
In The State of AI: How Organizations Are Rewiring to Capture Value, McKinsey & Company paints a picture of an AI revolution that has firmly taken root across industries. Yet most companies are still discovering what it takes to convert promising pilots into sustainable, enterprise-wide impacts. If there’s a unifying lesson, it’s this: AI success requires organizational change every bit as much as technical expertise.
By investing in workflow redesign, committing to careful risk management, training employees, and enlisting top-level leaders, businesses stand to unlock far greater returns from AI adoption. Although many are still in the early stages, the report’s findings suggest that the companies brave enough to undergo fundamental rewiring will reap the biggest rewards. For executives and managers looking to stay ahead of the curve, now is the time to take a thoughtful, strategic approach—one that lays a strong foundation for the future of AI-driven value creation.
Too Young To Be Cynical. Too Old To Be Naive.
4moVery nicely summarised, Indranil Sengupta, and your own comments helped put each of those findings in perspective. One small addition from my own baby steps in using gen AI to help in my work - under the few roles you mentioned under #6 above, I believe prompt engineering is a very important new skill that everyone should develop. I see evidence everyday that the lazier my prompt is, the more broad or even vague the output is. When I put some thought into writing a sharper prompt, I get better results.