McKinsey State of AI 2025: Agents, Innovation & Transformation Guide

What Is the McKinsey State of AI 2025 Report?

The McKinsey State of AI 2025 report, subtitled “Agents, Innovation, and Transformation,” provides one of the most comprehensive snapshots of artificial intelligence adoption across global enterprises. Based on a survey of 1,993 respondents from approximately 105 countries (38% from companies with over $1 billion revenue), this report reveals where AI stands in practice — not in press releases.

The paradox at the heart of the report is striking: 88% of organizations now use AI in at least one business function, up from 78% a year earlier, yet nearly two-thirds remain stuck in “experiment or pilot” mode. Only about a third have genuinely scaled AI across functions. The gap between AI ubiquity as a tool and AI impact as infrastructure defines the current moment.

For organizations navigating this transition, the McKinsey State of AI 2025 findings provide actionable benchmarks. Combined with insights from the McKinsey Global Institute research, the data paints a clear picture of what separates AI leaders from the rest.

AI Adoption: 88% Using, Few Scaling

The McKinsey State of AI 2025 confirms that AI has moved well beyond innovation labs. Over two-thirds of organizations now use AI in more than one function, spanning IT, marketing, customer operations, knowledge management, and beyond. If the question is simply “are you using AI?” — 2025 is the year of AI ubiquity.

But most adoption remains functional, not end-to-end. Among organizations using AI, approximately two-thirds stay in experiment or pilot mode, with only about a third reporting genuine scaling. Large enterprises ($5B+ revenue) are significantly more likely to have crossed the scaling threshold than smaller firms.

Three persistent blockers emerge: fragmented data and legacy technology, workflows designed for humans rather than AI, and a lack of clear scaling priorities. Organizations excel at “doing AI projects” but few know how to transform projects into a new operating baseline.

State of AI 2025: Agentic AI Adoption

AI agents — systems that can plan and execute multi-step workflows, not just answer questions — represent the frontier of the McKinsey State of AI 2025 analysis. At a high level, adoption seems promising: 23% of organizations say they are scaling agents in at least one function, while 39% are experimenting.

However, zooming into specific functions reveals a starker picture. In IT, software engineering, knowledge management, customer service, and marketing, no more than roughly 10% report agents as scaled or fully scaled in any single area. The gap between interest and impact is substantial.

The barrier isn’t technology but organizational readiness. To make agents useful, companies need standardized process steps, modern interfaces around legacy systems, and governance that makes autonomous action observable and interruptible. The EU AI Act is already establishing regulatory frameworks for autonomous AI systems that organizations must consider.

Only 6% Are True AI High Performers

McKinsey isolates a critical finding: only about 6% of organizations qualify as true AI high performers — those where more than 5% of EBIT is attributable to AI and leaders report significant value delivery. This tiny minority reveals what it takes to move from adoption to impact.

Their edge is overwhelmingly organizational, not technological. High performers are 3.6 times more likely to pursue transformational, enterprise-level change rather than incremental tweaks. They target growth and innovation with AI, not just cost savings. And they are far more willing to redesign workflows: 55% fundamentally reworked processes when deploying AI — nearly three times the rate of other firms.

This finding reinforces that AI success is fundamentally about organizational transformation. Plug-in thinking (adding AI tools to existing processes) consistently underperforms rewiring thinking (using AI as a catalyst to redesign processes entirely).

McKinsey State of AI 2025: ROI and Value Creation

The McKinsey State of AI 2025 report provides nuanced data on AI return on investment. While the majority of organizations report cost savings from AI deployments, high performers distinguish themselves by also capturing revenue growth — using AI to create new products, enter new markets, and enhance customer experiences.

The cost side of AI is growing. Organizations are increasing AI budgets substantially, with spending on talent, infrastructure, and model development all rising. However, the gap between AI spending and AI-attributable value remains wide for most organizations, suggesting that the challenge is less about investment levels and more about investment effectiveness.

For organizations benchmarking their AI programs, the WEF Future of Jobs Report provides complementary data on how AI is reshaping workforce structures and creating new value creation pathways across industries.

AI Talent and Workforce Transformation

The talent landscape in the McKinsey State of AI 2025 reflects the growing maturity of AI adoption. Demand for AI and machine learning engineers remains strong, but organizations increasingly recognize that technical talent alone is insufficient. AI product managers, AI ethicists, prompt engineers, and AI-human collaboration designers are emerging as critical roles.

High-performing organizations invest disproportionately in upskilling existing employees rather than relying solely on external hiring. They create learning programs that help non-technical staff understand and work with AI tools, building organization-wide AI literacy that enables broader adoption and more creative applications.

AI Governance as Competitive Moat

Perhaps the most underappreciated finding in the McKinsey State of AI 2025 is that governance is quietly becoming a competitive differentiator. Organizations with mature AI governance — clear policies, risk frameworks, monitoring systems, and accountability structures — are better positioned to scale AI with confidence.

Governance enables speed, not just safety. When teams know the boundaries within which they can operate, they move faster. When AI risks are systematically managed, organizations can deploy AI in more sensitive and higher-value domains. The NIST AI Risk Management Framework provides a structured approach that many organizations are adopting.

Industry-Specific AI Adoption Patterns

The McKinsey State of AI 2025 reveals significant industry variation in AI adoption. Technology, media, and telecommunications lead in AI agent deployment, with the technology sector already exceeding 90% AI usage. Healthcare and financial services show strong adoption driven by clear use cases in diagnostics, drug discovery, risk assessment, and customer service.

Manufacturing and energy sectors are accelerating, particularly in predictive maintenance, supply chain optimization, and process automation. The Tesla annual report and NVIDIA financials illustrate how AI leadership translates into competitive advantage in these sectors.

Implications for AI Strategy in 2026

The McKinsey State of AI 2025 delivers a clear strategic imperative: the window for experimental approaches to AI is closing. Organizations that haven’t begun scaling AI across functions risk falling irreversibly behind the 6% of high performers who are already capturing enterprise-level value.

Three strategic priorities emerge. First, shift from project-based AI to platform-based AI — building shared infrastructure and capabilities that multiple teams can leverage. Second, redesign work processes for AI-first operations rather than layering AI onto human-designed workflows. Third, invest in AI governance as a strategic capability that enables faster, safer scaling.

The full McKinsey State of AI 2025 report is available at McKinsey QuantumBlack.