AI's margin impact may be greatest in low-margin industries

I am increasingly convinced that the artificial intelligence opportunity is shifting away from model developers and vendors toward the end users. This is the RAMP framework I have been discussing since last year — real assets, margin potential.

Today there is fresh insight into how that might play out and what the upside could look like.

At this stage, you do not have to create the most advanced AI model to generate profit from the technology. A garbage truck fleet and a smarter routing system might be enough.

Waste Connections provides a compelling case study for where things may be going in hard-asset sectors. The company indicated to CIBC analysts that its initial wave of AI projects could produce about 100 basis points of margin expansion by 2030.

How will it get there?

"Pricing optimisation tools have already contributed to stronger retained pricing, while dynamic AI-enabled route optimisation is expected to improve fleet utilisation, labour productivity and routing efficiency in 2027," CIBC noted.

That serves as a template for other heavy-asset industries. Airlines have already talked about using AI to adjust fares, and I suspect more of that will happen as companies try to align prices with demand faster and capture the upside.

Customer service is another area where you have probably already seen the technology at work.

"WCN also highlighted customer-service automation as a significant opportunity, noting that the company receives ~700,000 to more than 1MM inbound customer calls per month and believes technology could ultimately eliminate 60%-65% of those interactions through proactive customer communication, mobile applications and agentic AI."

The company emphasised that this is only the start and it is already seeing quick returns on its technology investments.

How significant is 100 basis points of margin for earnings and the stock? In a sector like waste management, surprisingly not that much. The business is high margin already; shifting from 30% to 31% typically means a 4-5% lift to the share price. That is good but hardly a major endorsement of the huge capital spending going in and the sky-high multiples of model builders and their suppliers. Also, Waste Connections is a roughly $40 billion company that plans to spend only about $100 million on technology, and it is not clear how much is designated specifically for AI.

The area I would watch instead is low-margin industries: auto parts, aviation, steel, chemicals, packaging, food processing and trucking, for starters. At a company earning only a 3% operating margin, adding a single point lifts operating profit by a third. That is a target worth pursuing.