KTX ◆ Cattleman Capital · Market Read
Access Is Not Adoption: Why AI Early Movers Will Lead Construction Materials
Most producers can reach AI tools today. Very few have changed how they sell, price or operate because of them. That gap is where the next market leaders will come from.
Jason Cox, Consulting Division, KTX Cattleman Capital · 26-11-11 · As of 26-11-11 7:25 AM CT

I spend a lot of time with producers across the U.S. and Canada, and AI is no longer a mystery in those rooms. Somebody on the sales team is using a chatbot to draft emails. The dispatch or ERP vendor has added an AI feature. A few operators have bought enterprise licenses. Then I ask what changed on the P&L, and the room usually goes quiet.
I am talking to owners, CEOs, COOs, GMs and VPs of Sales in cement, ready-mix, aggregates, SCMs, precast, concrete products, import terminals and distribution. Almost everyone can get to these tools now. Very few use them well enough to move margin, volume or cash. The published data says most companies are stuck on access, and our industry is behind even that. The figures below come from published sources and look backward. Where I look ahead, that is my opinion.
McKinsey’s State of AI 2026 survey (published 26-08-25; 1,719 respondents) found that nearly nine in ten organizations regularly use AI in at least one business function. Only 37% attribute any EBIT impact to it, about the same as a year earlier. The “high performers,” those crediting AI with 5% or more of EBIT, held flat at about 6% of respondents (verified).
Look at that 6%. Nearly three-quarters of high performers say they fundamentally redesigned workflows around AI, against about one-quarter of everyone else (verified). They changed how the work gets done.
Dr. Kai-Fu Lee, author of the new book AI Native: The Mandate to Transform Your Company (Crown Currency, Penguin Random House), puts it bluntly in a TIME essay (26-11-03): most chief executives “can produce activity: pilots run, licenses purchased, employees trained, a partnership announced. None of that is a result.” Lee is chairman of Sinovation Ventures, founder and CEO of 01.AI, and a former senior executive at Apple, Microsoft and Google.
The U.S. Census Bureau’s Business Trends and Outlook Survey asks firms every two weeks whether they used AI in any business function. In the latest release (published 26-11-05, covering 26-10-05 to 26-10-18), 24.2% of all U.S. businesses said yes. Construction firms were at 17.1%, manufacturing (which includes cement and concrete products) at 21.4%, wholesale trade at 23.1%, and mining and quarrying at 10.2% (verified; single-period estimates move around).
The Census 2026 AI supplement asked AI users what they changed to make it work. 75.0% of construction firms using AI made no changes at all. Only 7.5% trained current staff and 10.0% built new workflows. Across all sectors, those figures were 64.3%, 15.0% and 15.4% (verified). Manufacturers did slightly better: 69.3% made no changes and 11.5% trained staff (verified). Among construction firms with no plans to use AI, 22.4% cited a lack of knowledge of what it can do (verified). That pattern is familiar. A new tool arrives, a handful of people try it, and the rest of the team goes back to selling the way they always have.
The sales desk is where this will show first
Sales and marketing is the one place where construction firms use AI most, and even there the numbers are thin: 7.7% of construction firms and 12.8% of manufacturers used AI in sales and marketing, against 14.3% of all businesses (Census 2026 AI supplement; verified). McKinsey respondents most often attribute AI revenue gains to marketing and sales (verified).
That matches what I see after a heavy bid week. Quotes waiting on follow-up pile up while the seller is on the phone with dispatch about a Friday pour, or rewriting a price letter because cement moved.
In a Gartner survey of 1,026 B2B sellers (published 24-09-16), sellers who effectively partner with AI tools were 3.7 times more likely to meet quota. The same survey found 50% of sellers felt overwhelmed by the amount of technology, and overwhelmed sellers were 45% less likely to hit quota (verified; 2024 data). Salesforce’s State of Sales 2026 (published 26-02-03; 4,050 sales professionals) found the average seller spends 40% of their time actually selling, and 51% of sales leaders with AI say disconnected systems are slowing their AI work (vendor survey; verified as published).
My read: drop more tools on a salesperson who already spends most of the day on everything except selling, and you will slow that person down. The edge goes to the team that picks three or four uses that fit its own market and runs them every day: account research before a contractor call, bid and quote follow-up, pricing and escalation language, territory and backlog reviews. AI can help with that prep. It does nothing if nobody has been trained to use it on live accounts.
A few producers are already doing the work
Heidelberg Materials reports that its Batch AI tool, built with Command Alkon, raised batch accuracy by up to 30 percentage points at ready-mix plants around Edmonton, with rollout continuing across North America (Heidelberg Materials; company-reported). Better accuracy means less cement overuse, which is the most expensive ingredient in the truck. Anyone who has stood on a batch deck knows what a high cement factor does to cost per yard.
Cemex built a sales copilot, Technical Xpert, that lets field sales reps pull product and technical answers in a chat window instead of chasing colleagues. It took four months to build and train on hundreds of pages of product documents (Microsoft feature, 24-04-03; vendor-published). A salesperson on a jobsite who needs a mix submittal or a product data sheet should not have to wait on a voicemail from technical services. Both started from a specific workflow with a measurable outcome.
Why I would not wait
In a Semafor interview (26-09-18), Lee described “a Darwinian choice” for CEOs. Digital businesses that fail to adapt will be “dead in the water,” he said, while businesses further from the AI frontier, such as coal mines, farms or manufacturing plants, may have a three-to-five-year “buffer” before they feel the pain of disruption (verified as reported by Semafor). He also proposed an “earnings call test”: if AI has not yet boosted profits, it is “AI theater.”
In a July essay on why he wrote the book, Lee argues the impact is unfolding over quarters rather than decades, and that the competitive gap compounds (X Article, 26-07-16; paraphrased). On 26-10-28 he wrote on X that white-collar workers and middle managers “who can neither work effectively with AI nor with people will face the greatest risk of displacement.”
My view: a heavy materials business may have that buffer. I would not treat it as time to sit still. Use the window to train people, rebuild a few core workflows and learn what works while competitors are still running pilots.
If I were sitting in your seat, I would start with the number: margin per yard or ton, quote-to-close rate, days sales outstanding, cement factor. If the AI work cannot point at a line on the P&L, it is a pilot. Train the sales and marketing team on your own customers, bids and territories. Skip the generic webinar. Have a manager coach it the way you coach a price letter. Redesign one workflow at a time, starting with quote follow-up after bid day, then pricing letters, backlog review or contractor prospecting. Measure before and after, and give each workflow an owner with the authority to change the process, not just the software.
The Consulting Division at KTX Cattleman Capital works with sales and marketing teams at cement, ready-mix, aggregate and SCM producers to decide which uses matter, train the team on live accounts, and tie the work to numbers leadership already tracks.
Innovate or Die
You have heard the saying “Innovate or Die.” The same holds true with artificial intelligence. Early movers in AI adoption will see their organizations transformed for the better. Those that hesitate, not even refuse but simply hesitate, will feel the impact.
At KTX we have a saying: “Tie Your Horse to a Solid Post.” In a market this uncertain, the solid post is a team that has been trained to use these tools well, not a stack of software licenses.
Sourcing notes
- Figures pulled 26-11-11, 7:05 to 7:25 AM CT.
- Census sector rates are single two-week estimates and carry sampling error. Census does not publish a separate rate for cement, ready-mix or aggregates; manufacturing (NAICS 31-33) and mining and quarrying (NAICS 21) are the closest published sectors.
- Dropped as unverifiable or weak: the MIT “95% of pilots fail” figure (preliminary, small sample), AGC/Sage 61% (primary not opened), and per-task agent speed and cost figures that appeared only in translation.
Sources
- https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
- https://www.census.gov/hfp/btos/data_downloads
- https://www.census.gov/hfp/btos/downloads/National.xlsx
- https://www.census.gov/hfp/btos/downloads/Sector.xlsx
- https://www.census.gov/hfp/btos/downloads/AI_Supplement_Table_2026.xlsx
- https://www.gartner.com/en/newsroom/press-releases/2024-09-16-gartner-sales-survey-reveals-sellers-who-partner-with-ai-re-three-point-seven-times-more-likely-to-meet-quota
- https://www.salesforce.com/news/stories/state-of-sales-report-announcement-2026/
- https://www.heidelbergmaterials.com/en/digital/batch-ai
- https://news.microsoft.com/source/latam/features/ai/cemex-technical-xpert-copilot/?lang=en
- https://time.com/article/2026/11/03/why-almost-every-corporate-ai-transformation-fails/
- https://www.semafor.com/article/09/18/2026/why-kai-fu-lee-thinks-companies-need-an-ai-boss
- https://x.com/kaifulee/status/2077911788491010275
- https://x.com/kaifulee/status/2115579574990037108
- https://www.ainativebook.com/
If a decision on your desk needs clearer context, send me a message.
Get the Cement & Concrete Week in Review by email: sourced reads for building materials producers in the U.S. and Canada. Subscribe on Substack
Jason Cox · Founder & Principal · Katy, Texas
First published on LinkedIn.
