[1/2] McKinsey’s Reports on Generative AI (Category AI)
Continuing my review of AI-adoption research, I liked McKinsey’s regular series on the subject. It began in 2023 and continued through 2024 and 2025, giving us enough history to examine trends over time. I will discuss these reports:
- 2023: The State of AI in 2023: Generative AI’s Breakout Year. Surveyed 1684 respondents on 11–21 April 2023; published 1 August 2023.
- Early 2024: The State of AI in Early 2024: Gen AI Adoption Spikes and Starts to Generate Value. Surveyed 1363 respondents from 22 February to 5 March 2024; published 30 May 2024.
- Mid-2024: The State of AI: How Organizations Are Rewiring to Capture Value. Surveyed 1491 respondents on 16–31 July 2024; published 12 March 2025.
- 2025: The State of AI in 2025: Agents, Innovation, and Transformation. Surveyed 1993 respondents from 25 June to 29 July; published 5 November 2025. I have already discussed its findings in more detail.
Some metrics across these reports can be read as time series.
❤️🔥 Adoption and its breadth
1) “AI use/adoption in at least one business function”: the main recurring KPI This is the most consistent metric in the series: the share of respondents whose companies use AI in at least one function.
- 2023: 55%.
- Early 2024: 72%.
- Mid-2024: 78%.
- 2025: 88%. It measures penetration—whether AI is used anywhere—but does not distinguish pilots from production particularly well.
2) “Organizations regularly using gen AI in ≥1 business function”: another recurring KPI The November 2025 edition does not emphasize this measure.
- 2023: 33%.
- Early 2024: 65%.
- Mid-2024: 71%.
- 2025: N/A. The text gives no separate figure, focusing instead on agents and AI overall.
3) “AI spreads across multiple functions”: breadth of use, only partly comparable This one is trickier. The idea stays similar, but the wording changes: ≥2 functions, more than 1 function, or ≥3 functions.
- 2023: less than 33% had adopted AI in two or more functions.
- Early 2024: 50% had adopted it in two or more functions.
- Mid-2024: N/A.
- 2025: more than 66% used AI in more than one function, and 50% in three or more functions. The authors seem more interested in finding ways to measure breadth than in year-to-year comparability, gradually raising the number of functions in the question.
In the next part, I will discuss AI and GenAI’s reported effects on companies, and the “high performers” who get more from adoption. I have covered that here, but the continuation shows how the high-performance criteria change over time.
#Engineering #AI #Metrics #Software #Productivity #Economics #Whitepaper