Australian small business · AI & digital index · v2

Every number verified against its source. Every month.

What is actually happening where Australian small business meets AI: who is using it, how AI answers are changing search, what data breaches look like, how small businesses are trading. Each figure below carries the exact line from its public source, and a script re-checks that line against the live page every month. Where the published numbers disagree, the first table says why.

9 indicators 6 sources 4 figures reconciled Last verified 1 October 2026

The question everyone asks

Which Australian AI-adoption figure is right: 12%, 42%, two-thirds or 58%?

All of them, for what each one measured. The Bureau of Statistics counts every Australian business over a financial year and gets 12%. NAB surveys small and medium businesses in 2026 and gets 42%. Deloitte surveys a thousand SMBs and finds two-thirds using AI at all, with 5% using it well. Roy Morgan asks individuals, not businesses, and gets 58%. The spread is the difference in who was asked and what counted, not a disagreement about the facts. Quote the one whose population matches your sentence.

12%
Who was measured
All Australian businesses (ABS Business Characteristics Survey, 2024–25)
What counted
Reported use of artificial intelligence in the business during the year
verbatim ▸

“Just over one in ten (12%) businesses reported use of Artificial Intelligence (AI), compared to 1% in 2022–23”

42%
Who was measured
Small and medium businesses surveyed by NAB with MYOB and Indeed
What counted
Currently using AI
verbatim ▸

“NAB says 42% of Aussie SMEs use AI to boost productivity”

Two-thirds; 5% fully enabled
Who was measured
More than 1,000 Australian SMBs surveyed for Deloitte's AI Maturity Index
What counted
Using AI at all, versus using it with the strategy, data and skills to realise its benefits
verbatim ▸

“while two-thirds of SMBs are using AI, just 5% of surveyed SMBs using the technology are fully enabled to realise its potential benefits”

58%
Who was measured
Australians aged 14 and over (individuals, not businesses)
What counted
Used an AI tool in an average four-week period
Mar 2026 Roy Morgan
verbatim ▸

“13.6 million people, equivalent to 58% of Australians aged 14+, use ‘Artificial Intelligence’ (AI) tools in an average four weeks”

AI adoption

58%
as of Mar 2026
Australians (14+) using AI tools in an average four weeks
13.6 million people, March quarter 2026
Source Roy Morgan
verbatim from the source ▸

“13.6 million people, equivalent to 58% of Australians aged 14+, use ‘Artificial Intelligence’ (AI) tools in an average four weeks”

Verified 2026-10-01

45%
as of Mar 2026
Australians using ChatGPT, the most-used AI tool
10.5 million Australians, March quarter 2026
Source Roy Morgan
verbatim from the source ▸

“10.5 million Australians (45%) using ChatGPT”

Verified 2026-10-01

AI search

58%
as of Apr 2026
Lower average clickthrough for the top-ranking page when an AI Overview is present
Google Search Console data across 300,000 keywords
Source Ahrefs
verbatim from the source ▸

“the presence of an AI Overview now correlates with a 58% lower average clickthrough rate for the top-ranking page”

Verified 2026-10-01

8%
as of Mar 2025
Share of Google visits with an AI summary where users clicked a traditional result
vs 15% on pages without an AI summary, March 2025 browsing data
Source Pew Research Center
verbatim from the source ▸

“Users who encountered an AI summary clicked on a traditional search result link in 8% of all visits”

Verified 2026-10-01

68.01%
as of Apr 2026
Google searches ending without a click (United States)
January–April 2026, Similarweb clickstream panel. US data. No equivalent Australian panel is published, so read it as directional for Australia rather than local.
Source SparkToro / Similarweb
verbatim from the source ▸

“In the first four months of 2026, a whopping 68.01% of Google searches ended without a click”

Verified 2026-10-01

Data security

532
as of June 2025
Notifiable data breaches reported to the OAIC, January–June 2025
a 10% decrease on the previous six months
verbatim from the source ▸

“In the January–June 2025 reporting period, we received 532 data breach notifications, a 10% decrease compared with the previous 6 months”

Verified 2026-10-01

18%
as of June 2025
Share of breach notifications from the health sector, the most-breached sector
of 532 notifications, January–June 2025
verbatim from the source ▸

“The health sector had the most reported data breaches (18% of reported data breaches)”

Verified 2026-10-01

Small business economy

6.5%
as of June 2026
Australian small business sales growth, year on year
June quarter 2026
Source Xero Small Business Insights
verbatim from the source ▸

“Sales grew a still healthy 6.5% year-on-year”

Verified 2026-10-01

22.9 days
as of June 2026
Average days an Australian small business waits to be paid
June quarter 2026
Source Xero Small Business Insights
verbatim from the source ▸

“The average length of time small businesses waited to be paid, after issuing an invoice, was 22.9 days”

Verified 2026-10-01

What changed

  1. 2026-09-23 Index v2: history per indicator, the AI-adoption reconciliation table, CSV and JSON downloads, CC BY 4.0 licence, Dataset markup completed.
  2. 2026-09-22 Two OAIC indicators moved to the January–June 2025 release; US zero-click figure added and labelled US-only; 18-month age check introduced.
  3. 2026-08-03 Index published with source-verified indicators from Roy Morgan, the OAIC, Xero, Ahrefs, Pew Research Center and SparkToro.

Methodology

The index curates published figures from public Australian and international sources: Roy Morgan, OAIC Notifiable Data Breaches statistics, Xero Small Business Insights, Ahrefs, Pew Research Center, SparkToro / Similarweb, plus the reconciled adoption studies above. We do not run surveys and we do not model estimates; we curate, attribute and verify.

Verification is mechanical. Each figure stores the exact sentence or table line it came from. A deterministic script re-fetches every source monthly and string-matches the stored line against the live page. If a source changes, the figure is flagged and a person reviews it before any value changes. A figure older than 18 months is flagged for replacement even when it still verifies. A figure that can no longer be verified is removed rather than left to drift, and the change is logged above.

When a value changes, the previous value stays in the indicator's history, so the page shows movement over time rather than a snapshot. The CSV and JSON are generated from the same verified file the page is rendered from; they cannot disagree with it.

Anchors are permanent. A figure's URL (#ai-adoption-au and so on) will not move; if an indicator is retired, its anchor stays and says so.

If we are not the right people to maintain a number, we will say so and point at whoever is. Corrections: tell us what we got wrong and we will fix it, with the change noted.

Cite and download

Published under Creative Commons Attribution 4.0. Cite the original source for any underlying figure. If the aggregation, reconciliation or verification is what you are using, cite the index; every figure above has a "Copy citation" button and a permanent anchor.

Compound Co, The Australian Small Business AI & Digital Index, v2, verified 1 October 2026.
https://www.compound-co.com.au/australian-smb-ai-index/

Frequently asked questions

Which Australian AI-adoption figure should I use: 12%, 42% or 58%?

The one whose population matches your sentence. The ABS figure (12%) covers all Australian businesses reporting AI use in 2024–25. NAB's 42% is small and medium businesses currently using AI, surveyed in 2026. Roy Morgan's 58% is individual Australians aged 14 and over, not businesses. Deloitte's two-thirds is surveyed SMBs using AI at all, of which only 5% were fully enabled. They do not contradict each other; they measure different things.

How is the index verified?

Every figure stores a verbatim quote from its public source. A deterministic script re-fetches each source monthly and string-matches the quote against the live page, the same mechanism that verifies citations on our blog. A figure whose source changes is flagged and reviewed by a person before the value is updated; a figure older than 18 months is flagged for replacement; nothing on this page is machine-invented.

Can I cite, download or republish these figures?

Yes. The index is published under CC BY 4.0: use it with attribution. Download the CSV or JSON, cite the original source for the underlying figure, and cite or link this index if the aggregation, reconciliation or verification is what you are using. Each figure has a permanent anchor.

How often is the index updated?

Sources are re-verified monthly by machine. Values change only when a source publishes new data (quarterly for most of the underlying series), and every change is reviewed by a person and recorded in the indicator's history and in the change log on this page.