# Sources for our figures

> Every figure cited on aiorigin.io, with its publisher, date, sample and a link to the source. AI Origin’s own calculations are flagged as such.

URL: https://aiorigin.io/en/sources

- Home
- Resources

# Sources for our figures.

Every figure shown on the site, with its source, date and link. Our own calculations are flagged. Updated: October 13, 2026.

Contents
01The figures on the home page02AI capability and adoption in Europe03Reference figures on AI blockers04Figures the site no longer shows.05Figures left out.06Simulations.

## The figures on the home page

The two figures in the home page sequence come from McKinsey’s annual global survey.

- 88 companies out of 100 use AI in at least one function (2025 edition, published 5 November 2025; the 2026 edition says “nearly nine in ten”).
- 6 out of 100 attribute at least 5% of their operating profit to AI and describe its value as significant: McKinsey calls them AI high performers (2026 edition, unchanged since 2025).

McKinsey, The state of AI in 2026: On the road to ROI, 25 August 2026. mckinsey.com (new tab) ↗

McKinsey, The state of AI in 2025: Agents, innovation, and transformation, 5 November 2025.

## AI capability and adoption in Europe

×180

“×180. Since 2023, AI can complete tasks 180 times longer on its own.”

Source

METR, Time Horizon 1.1. Latest measurement added on May 8, 2026; since September 8, 2026, METR has stated that the page is no longer updated.

Type

Benchmark measurement, published by an independent research lab.

Scope

50% task-completion horizon of frontier AI models at release, on software and research tasks. GPT-4 (March 2023): 3.99 minutes. o3 (April 2025): 119.7 minutes. Claude Opus 4.6 (February 2026): 718.8 minutes, or about 12 hours, the highest horizon measured for a publicly available model according to METR. A preliminary version of Claude Mythos Preview (April 2026) reaches 1,044.8 minutes (≈ 17 h 25 min; confidence interval: 8 h 29 min to 55 h 04 min).

Method

×180 = 718.8 ÷ 3.99 (AI Origin calculation). Doubling time: 128.7 days since 2023 according to METR (confidence interval: 104 to 158 days), or about 4 months. For models released since 2024 alone, METR measures a doubling time of 105 days (Frontier Risk Report, May 2026).

Caveat

This is the length, in human expert time, of the tasks an AI completes successfully half the time. Not an AI that “works for 12 hours”. METR considers its measurements beyond 16 hours unreliable. No public model has reliably surpassed Opus 4.6: GPT-5.4 and Gemini 3.1 Pro, measured after it, remain below 6 h 30 min. The Mythos Preview data point exceeds that threshold, and the model is only available to Anthropic partners. Models released since then (Claude Opus 5.5, GPT-5.6 Sol…) have no reliable measurement: the task suite is practically saturated.

Links

metr.org/time-horizons (new tab)Raw data (YAML) (new tab)Frontier Risk Report (new tab)GPT-5.6 Sol evaluation (new tab)Claude Opus 5.5 evaluation (new tab)Anthropic, Project Glasswing (restricted access to Mythos Preview) (new tab)

8% 20%

“×2.5. In two years, the share of European companies using AI.”

Source

Eurostat, news item of December 11, 2025.

Scope

Companies in the European Union with 10 or more employees using at least one AI technology: 8.1% in 2023, 13.5% in 2024, 20.0% in 2025.

Method

×2.5 = 20.0 ÷ 8.1, rounded (AI Origin calculation).

Link

ec.europa.eu/eurostat (new tab)

≈ 31%

“EU 2026 · 31%”

Source

AI Origin calculation based on the 2023-2025 Eurostat series (Eurostat entry).

Method

Average annual rate for 2023-2025, extended by one year: 20.0% × (20.0 ÷ 8.1)½ ≈ 31.4%, index ≈ 388. This is the same assumption as doubling in ≈ 18 months. Other methods give between 26% and 32%.

Caveat

This is not a measurement. Eurostat plans to publish the 2026 figures in December 2026; they will replace this projection. The chart shows the rounded value, 31%; the source line below the figures flags it as an AI Origin calculation.

7.5%

“Only 7.5% of companies get value from it at scale.”

Source

BCG, Applied AI Index 2026 (report “The Formula for Agentic AI Value”), press release of September 30, 2026.

Scope

1,330 executives. The 7.5% are the companies BCG describes as “future-built”.

Links

BCG press release (new tab)Report (new tab)

## Reference figures on AI blockers

71%

“71% of French companies that do not use AI say they see no use for it.”

Source

Insee, Insee Première n° 2120, July 21, 2026 (2025 data).

Scope

Companies with 10 or more employees that do not use AI, excluding the financial and agricultural sectors.

Link

insee.fr/fr/statistiques/9025878 (new tab)

48%

“48% of employees who use AI at work have already entered company information into public tools.”

Source

KPMG and the University of Melbourne, Trust, attitudes and use of artificial intelligence: A global study 2025.

Scope

48,340 people in 47 countries, surveyed from November 2024 to January 2025. Base: employees who use AI at work.

Link

Report (PDF) (new tab)

49%

“In France, 49% of users have not told their manager.”

Source

Ifop for Talan, barometer “Les Français et les IA génératives”, March 2025.

Scope

1,100 people. The figure covers professional users of generative AI who have a line manager, about 17% of the sample.

Caveat

Small sample: read it as an order of magnitude. On the site, it always follows the KPMG figure, measured on a much larger sample.

Link

Ifop report (PDF) (new tab)

≈ 2 in 3

“Nearly two-thirds of organizations have not yet begun scaling AI across the enterprise.”

Source

McKinsey, The state of AI in 2025: Agents, innovation, and transformation, November 5, 2025.

Scope

1,993 respondents in 105 countries, survey conducted from June 25 to July 29, 2025.

Link

mckinsey.com (new tab)

72%

“72% of companies cite integrating AI with their systems, tools and APIs as a major challenge.”

Source

BCG, The Widening AI Value Gap, September 2025 (exhibit 10: challenges cited by respondents).

Scope

More than 1,250 companies worldwide.

Link

Report (PDF) (new tab)

42% · 5%

“42% of companies want to give AI agents real autonomy by 2030.”

Source

BCG, Applied AI Index 2026, press release of September 30, 2026.

Scope

1,330 executives.

Links

BCG press release (new tab)Report (new tab)

72% · 36%

“72% of employees say AI has upended the skills expected of them.”

Source

BCG, AI at Work: Strategy Matters More Than Tools, June 3, 2026.

Scope

11,749 respondents in 14 markets.

Links

BCG publication (new tab)Press release (new tab)

42% · 66%

“42% of regular frontline users save at least one day a week.”

Source

BCG, AI at Work: Strategy Matters More Than Tools, June 3, 2026.

Scope

11,749 respondents in 14 markets. The 42% refers to frontline employees who use AI regularly.

Links

BCG publication (new tab)Press release (new tab)

## Figures the site no longer shows.

Cited in the previous version of the site. Their sources remain here.

- 47.8% of working-age French people use generative AI.: Microsoft AI Economy Institute, AI Diffusion Report, first quarter of 2026, published on May 7, 2026. Vendor telemetry: it measures people, not companies. news.microsoft.com (new tab)
- The share of companies using AI doubles roughly every 18 months.: AI Origin calculation on the Eurostat series: 24 months × ln 2 ÷ ln (20.0 ÷ 8.1) ≈ 18.4 months (about 15 months on the Insee series). An order of magnitude, not a forecast.
- In France, the share of companies using AI rose from 6% in 2023 to 18% in 2025.: Insee, Insee Première n° 2120, July 21, 2026. Companies with 10 or more employees, mainly market-sector, excluding the financial and agricultural sectors: 6% in 2023, 10% in 2024, 18% in 2025. We never use this figure to talk about banking or insurance. insee.fr (new tab)
- 55% of executives cite redesigning organizations and jobs as a major barrier.: BCG, press release of July 22, 2026. 152 CEOs of companies with at least $500 million in revenue: a small sample. BCG press release (new tab)
- Companies that get value from AI at scale show revenue growth 2.4 times higher than laggards.: BCG, Applied AI Index 2026, press release of September 30, 2026. 1,330 executives. BCG press release (new tab)
- 76% of French people say they have concerns about generative AI; 73% feel they lack the knowledge to use it.: Ifop for Talan, barometer “Les Français et les IA génératives”, March 2025. 1,100 people surveyed in France. Ifop report (PDF) (new tab)

## Figures left out.

Often cited, they do not appear on the site. Here is why.

- MIT NANDA, “95% of pilots fail”: Preliminary study with a disputed methodology.
- “79% are experimenting, fewer than 10% have scaled”: Original source not identified.
- BCG 2025, “60% of companies get no value”: Superseded by the 2026 edition of the study.
- McKinsey 2026, “80% more productive, 37% impact on earnings”: Original report not reviewed.
- Gartner, “over 40% of agentic AI projects canceled by 2027”: A forecast, not a measurement.

## Simulations.

Scenes and figures marked “Simulation” on the site use demo data. They do not describe any client.

Write to us at contact@aiorigin.io: we correct it, and the update date of this page changes with the correction.

## Let’s find where to start, together.

30 minutes with our team to spot what is slowing yours down. No commitment, no sales pitch.

contact@aiorigin.io

Thank you, your request has arrived.
We will reply to set up a time.
