The State of AI in 2025

Three Questions McKinsey’s Report Answers

The latest McKinsey Global Survey on AI & Gen AI, which forms part of the State of AI survey report 2025, finds that organizations are beginning to take steps that drive bottom-line impact. Here are the 3 questions this report answers:

🛠️ 1. How are leading companies rewiring themselves to capture Gen AI value?

CEO ownership is decisive. Among 25 organizational traits, the one most strongly linked to higher EBIT from Gen AI is when the CEO personally oversees AI governance .
Workflow redesign beats pilot projects. A fundamental re-thinking of processes—not bolt-on tools—drives the biggest bottom-line lift. Yet only 21 % of Gen AI adopters have rebuilt any workflows so far .
Large enterprises are moving first. Firms above $500 M revenue centralize data-risk oversight and hire faster for AI talent, giving them a head-start on value capture .

Story in a sentence: McKinsey’s 2025 Survey Report on AI & Gen AI reveals that winners treat Gen AI as a C-suite-led transformation, not an IT experiment. Read McKinsey’s Seizing the Agentic AI Advantage Report to see how CXO’s at leading firms are rewiring themselves.

🚀 2. What practices turn Gen AI pilots into scaled impact?

The report tests 12 adoption-and-scaling habits and finds two stand out:

High-impact habit	Reality check
Track clear KPIs for every Gen-AI solution	< 20 % of companies do it today
Publish a phased road-map and dedicated adoption team	Practiced by > 40 % of large firms, but barely 17 % of smaller ones

According to McKinsey’s 2025 Survey Report on AI & Gen AI, less than one-third of respondents follow even “most” of the 12 practices, explaining why enterprise-level profit gains remain rare.

⚠️ 3. Where are early Gen AI implementation results—and risks—showing up?

Revenue & cost: Business units that embedded Gen-AI now see rising sales (e.g., 70 % of product-development users report revenue bumps) and majority cost cuts across supply-chain, service, and IT functions .
Risk management matures: Mitigation of inaccuracy, IP, and privacy threats jumped 10-20 pts vs. early 2024, led by big companies .
Talent & workforce: Half of adopters expect to hire more data scientists next year; 44 % have already reskilled up to 10 % of staff, and plan to double that within three years . Head-count is predicted to fall in service ops but rise in product and software engineering .

Net-net: Value is finally visible inside functions—but enterprise EBIT will lag until KPI discipline, risk guardrails, and workforce scaling catch up.

The bottom line

Gen-AI’s frontier has shifted from experimentation to enterprise rewiring. Put the CEO in charge, redesign at least one high-value workflow end-to-end, measure ruthlessly, and fund the talent and risk muscle that make AI scalable and safe. Organizations that act now convert curiosity into competitive cash flow; those that wait stay stuck in pilot purgatory.

Read the McKinsey report

McKinsey