Connecting: 216.73.217.151
Forwarded: 216.73.217.151, 172.71.28.133:33707
Using AI in ‘real world America’ | Trustnet Skip to the content

Using AI in ‘real world America’

17 November 2026

For economic value to keep accruing to businesses using AI, they need to see 'deflationary token economics' continue.

By Phil Coombes

De Lisle Partners

AI models such as Claude and ChatGPT can now do much more than suggest answers to emails, summarise internet searches or act as glorified chatbots – and the gains are starting to show up in real-world businesses.

The financial media tends to focus on the AI technology companies themselves, discussing whether we are building too many data centres or what constitutes fair value for Anthropic as it seeks a $1.4trn blockbuster IPO.

But alongside these headlines, the underlying AI models have been improving rapidly, aided by fierce competition between the leading AI labs: Anthropic and OpenAI, and increasingly Google, Meta, SpaceX (which owns xAI, maker of Grok), as well as a series of Chinese players. Users are now seeing the benefits of far more powerful models, with greater usage allowances for similarly priced subscriptions.

AI models can now automate entire workflows, analysing and reshaping complex data sets at speed and erasing tedious tasks from to-do lists. Companies deploying AI automation at scale are capturing genuine productivity gains across their businesses, boosting sales, reducing costs, increasing innovation and growing margins.

This creates opportunities to identify AI beneficiaries well beyond the technology sector and Silicon Valley.

 

AI in practice

While much has been written about how AI can help banks, investment teams and accountancy firms save time and money, all kinds of US companies are seeing the benefits of using AI agents to process tasks.

Pennant Group, a home healthcare provider, is piloting AI tools to streamline clinicians’ paperwork and automate back-office processes. Customers Bancorp, a tech-led regional bank, is cutting commercial loan closing times with an agentic AI tool.

Valmont, an infrastructure equipment manufacturer, is deploying AI-enabled scheduling tools in its factories to lift throughput. And Titan America, a cement and concrete producer, is using AI for predictive plant maintenance and streamlining its logistics.

The results are now beginning to show up in quarterly earnings reports.

For example, Worthington Enterprises, which manufactures building products, gas canisters and professional tools for industrial and consumer markets across North America, has deployed AI across its entire business to increase efficiency.

In its fourth quarter 2026 earnings call, Worthington's CEO called out the ways in which AI deployment had helped overheads shrink as a share of sales from 22.5% to 21% compared with the previous year.

This was achieved by implementing its AI-aided 'Worthington Business System' to produce incremental gains right across the business and has led to revenue growth outpacing selling, general and administrative (SG&A) costs for six consecutive quarters.

 

Cheap inference is here to stay

For economic value to keep accruing to businesses using AI, they need to see 'deflationary token economics' continue, that is, a stable budget buying an ever-increasing amount of AI work.

We have all borne witness to tech company 'shrinkflation', where the reverse plays out: more adverts, or less service, for the same subscription. Think Netflix and Spotify. Could Anthropic tighten the screws after its IPO and erode the value now accruing to real-world America?

Nothing is certain, but a key dynamic shaping the AI race has been competition from open-source models, initially from Chinese AI labs, now from nascent Western alternatives. These open models are offering comparable performance at a much lower cost.

Developers are now using the more expensive Claude and ChatGPT models, with their higher intelligence, to design their workflows, but then outsourcing the recurring running of that work to cheaper models.

The AI model aggregator OpenRouter, which allows access to multiple models in one place, reports that more than 70% of tokens consumed via its platform now come from its clients using open source models.

While competition in developing AI models remains intense, much of the economic value can keep accruing to the businesses that keep up with developments and implement AI effectively.

There is little doubt that those able to glean productivity gains from automated workflows stand to become more competitive and expand their margins. The real AI winners may sit far from Silicon Valley, in unassuming offices across middle America.

Phil Coombes is an investment analyst at De Lisle Partners. The views expressed above should not be taken as investment advice.

Editor's Picks

Loading...

Data provided by FE fundinfo. Care has been taken to ensure that the information is correct, but FE fundinfo neither warrants, represents nor guarantees the contents of information, nor does it accept any responsibility for errors, inaccuracies, omissions or any inconsistencies herein. Past performance does not predict future performance, it should not be the main or sole reason for making an investment decision. The value of investments and any income from them can fall as well as rise.