Sector-by-sector breakdown of AI automation adoption, with per-sector sample sizes. Every figure is an unfilled placeholder until it carries a citable, dated source.
2 min read
The structure of our annual read on AI automation in production. Every figure is an unfilled placeholder pending sourced research — nothing here is publishable yet.
2 min read
Every figure below is an unfilled placeholder. Do not fill any of them from memory or inference; each requires a citable source with a date, or the row is removed.
Aggregate adoption numbers hide the only thing a buyer wants to know: whether organisations that look like theirs, with their regulatory load and their data, have made this work. This piece splits the same underlying survey by sector and reports each one with its own sample size, so a sector with too few responses can be seen to be too small to trust.
For each sector: production adoption rate, most common first workflow, median time to first production deployment, and the most-cited blocker.
Per-sector sample sizes: [PLACEHOLDER — verify before publishing]. Any sector below the minimum reportable sample is to be shown as "insufficient sample" rather than given a number.
Hypotheses to be tested against the data rather than published as conclusions: process standardisation, the existence of a machine-readable document trail, regulatory tolerance for automated decisions, and in-house engineering capacity. Supporting figures: [PLACEHOLDER — verify before publishing]
Each sector's constraints on automated decision-making differ, and they change. This section cites the applicable rules and their dates rather than characterising them from memory: [PLACEHOLDER — verify before publishing]
Guidance to be written once the sector cuts are final.