More helpers than nails: 84% of organizations in Kahua's survey use general-purpose AI tools. Only 8% report significant measurable value.

You decide to hang one picture. Someone fetches a hammer. Someone else fetches a tape measure. A neighbor turns up with a level. Before long there are more helpers than nails, and every one of them is sure of a different spot on the wall. Nobody is short of tools. The picture is still on the floor.

Kahua's new survey finds the same thing with AI. In its release of The 2026 Capital Program Maturity Report, 84% of surveyed organizations use general-purpose AI tools, and only 8% (4 of the 50) report significant measurable value from their AI investments. The sample is 50 capital program leaders across the U.S., U.K. and Canada.

As for where the mark goes, only 16% are very confident in the accuracy and timeliness of their project and portfolio data. Only 4% operate in a fully connected systems environment.

Kahua sells an enterprise construction platform it pitches as a connected system of record, and it commissioned the survey. BNN Bloomberg put the same gap on air on 23 October 2026 with Kahua's CEO. That gap is the story: adoption is not value.

The report was commissioned by Kahua and researched by Verdantix. Kahua's public pages date the release to October 2026. The news dateline says 14 Oct, the news page and PR Newswire say 20 Oct, and a Kahua trends article dated 9 Oct already cites the finding. It draws on 50 quantitative survey interviews and five in-depth interviews with capital program leaders across the U.S., U.K. and Canada. Kahua's report page says the firm partnered with Verdantix "to survey 50 capital program leaders to benchmark the digital maturity of their construction programs." Kahua sells an AI enterprise construction platform and describes itself as a system of record for owners and delivery teams. That gives it a commercial interest in the finding that connected systems and native AI beat side tools.

The wording of the AI finding shifts slightly across Kahua's own pages. The news release and the PR Newswire copy say organizations "use general-purpose AI tools" and "report significant measurable value." The report landing page says "84% use public-based AI models like ChatGPT, but only 8% are getting much value from it." Kahua's construction trends article is close to the news-release wording but drops "tools" ("general-purpose AI"). BNN's host said "nearly 84 per cent" and "actual measurable values," and BNN's summary said "significant value." AIVI uses Kahua's news-release wording for the headline figures and treats the other phrasings as the same survey, not a second measurement.

What the public materials do publish about the gap points at data and systems, not at headcount. Only 4% of surveyed organizations operate in a fully connected systems environment, and just 16% are very confident in the accuracy and timeliness of their project and portfolio data. Only 22% currently have a system of record spanning the full project lifecycle, while 58% say they are moving toward reducing the number of systems they operate or consolidating onto a unified platform. Kahua's landing page also says 94% of organizations are managing six or more projects at once. None of those figures is a productivity baseline for AI. They are the conditions Kahua and Verdantix put next to the 8%.

On value, the respondents' own expectations are about speed and productivity. Kahua's report page says they "expected AI to improve speed, insight, and productivity, not accountability, which remains a human function." In the BNN interview, CEO Scott Unger described AI as a productivity tool for defined construction processes such as punch lists, submittals and reading asset nameplates, and said AI "still certainly requires the human in the loop." No source in the release or the interview claims headcount reductions. For AIVI, value means a productivity lift you can check. This survey reports self-assessed "significant measurable value" and doesn't say what that meant or how it was measured.

Sophie Planken-Bichler, industry analyst at Verdantix, said in Kahua's release: "The research highlights the difference between having access to AI and actually putting it to work." Unger, on Kahua's side, said technology alone doesn't solve problems and that a trusted system of record is the foundation for native AI to deliver meaningful value. Both are arguing Kahua's thesis. BNN did not independently re-run the survey.

What would make this a stronger Index signal: Verdantix or Kahua publishing the question that produced the 8%, the definition of significant measurable value, fieldwork dates, response rate and margin of error, and any productivity unit behind "measurable." Until then, it sits with other vendor surveys as unscored adoption research. It's useful as a caution that use is not value, not as a graded deployment.