Skip to content
back to the archive page
#AI

[1/2] The State of AI in 2025: Main Findings (Category AI)

#AI #Engineering #Metrics #Software #Productivity #Economics #Whitepaper

[1/2] The State of AI in 2025: Agents, Innovation, and Transformation — Main Findings (Category #AI)

I studied an interesting November 2025 report prepared by QuantumBlack, AI by McKinsey, under a group of McKinsey senior partners. I wanted to compare its findings with the joint report by Yakov and Partners, formerly McKinsey, and Yandex that I covered earlier.

McKinsey’s global study is based on an online survey conducted from 25 June to 29 July 2025. It included 1 993 respondents from 105 countries, spanning major regions, industries, company sizes, functions and seniority levels. Around 38% work at companies with more than $1 billion in revenue. Responses were weighted by countries’ shares of global GDP to reduce imbalances.

Key definitions used throughout the report:

  1. “Regular AI use” means use in at least one business function. Adoption is divided into stages:
  • Experimentation.
  • Piloting.
  • Scaling, or deploying solutions across the company.
  • Full enterprise-wide scaling.
  1. “AI agents” are foundation-model-based systems that can act autonomously in the real world, planning and executing multistep tasks.
  2. “AI high performers” are a small group, around 6% of surveyed organizations, where AI directly accounts for more than 5% of total company EBIT and “significant” business value. Their practices are examined separately.

The executive summary opens the report with these points:

1️⃣ Most organizations are still experimenting or piloting Almost two-thirds of respondents say their companies have not yet begun scaling AI enterprise-wide.

2️⃣ Interest in AI agents is high 62% say their organizations are at least experimenting with agents.

3️⃣ There are early positive signs, but not yet across the whole business Respondents see lower costs and higher revenue in specific use cases; 64% believe AI supports innovation. Only 39%, however, report an impact on EBIT at the company-wide level.

4️⃣ High performers use AI for growth and innovation as well as efficiency 80% say their companies target efficiency gains in AI initiatives. Those obtaining the most value often also pursue growth or innovation rather than limiting themselves to cost cuts.

5️⃣ Workflow redesign is a key success factor Half of the high-performing companies plan to use AI to transform the business, and most are already redesigning workflows and task sequences.

6️⃣ Expectations for employment differ For the coming year, 32% expect AI to reduce headcount, 43% expect no change and 13% expect growth.

In the next post, I’ll discuss the strategies of high-performing ~~managers~~ teams that the report treats as a golden image: a model for those not yet ready to report an AI contribution of 5% of total EBIT.

#Engineering #AI #Metrics #Software #Productivity #Economics #Whitepaper