AI is rewriting how global business runs. Which parts earn their keep?

You saw it in the last forecast that missed, the supply disruption that arrived without warning, the market you entered that never quite performed. And you have started to notice something about the competitors who seem to see these things coming: they are not simply luckier or braver. They are better informed, earlier, and increasingly it is not a person doing the informing. The quiet worry is not that you are being outspent. It is that you might be getting out-predicted.

Here is where that leads. A 2024 University of the Sunshine Coast review in the Thunderbird International Business Review, which synthesised 37 studies on AI in international business, maps where the technology genuinely earns its keep. The standout is operations. It cites research in which artificial neural networks (software modelled loosely on the brain, trained on many past scenarios to predict new ones) forecast supply-chain risk with 96 to 100% accuracy, and an AI-optimised supply model that raised profit by around 30% while holding inventory at its lowest. The pattern is clear: AI pays best where the decision is data-rich and repeats. Point it there first.

Where does AI actually earn its keep in global operations?

In the supply chain, ahead of almost everywhere else. The review gathers the evidence plainly: beyond that 96 to 100% risk-prediction accuracy, AI is being used for demand forecasting, inventory and route optimisation, production scheduling, logistics planning, and the monitoring of supplier quality, delivery and price. One cited study built an AI-enhanced model that lifted profit by roughly 30% while minimising stock. Smart factories increasingly run autonomously, and AI is used to steer global value chains (the international webs of suppliers and plants behind a finished product) by predicting trends and protecting on-time delivery.

This is the same shift reshaping the engine of business growth itself: the repeatable, data-heavy work is being industrialised first. An operational decision made daily, on rich data, with a clear measure of success, is exactly the kind of work AI does well, which is why the supply chain is where the early advantage is concentrating.

What about choosing markets and entering them?

Here AI shifts from running the work to sharpening the judgement around it. The review describes AI predicting which international markets are most attractive, comparing the cultural and psychological distance between countries, and generating country, industry and competitor profiles in minutes using generative tools. There are models built specifically to help smaller firms with limited resources choose markets, and machine translation that has measurably opened cross-border sales, as it did for sellers on eBay.

One caution the review draws out is worth carrying. When you enter a country, weigh its digital capability honestly, because it can either carry your AI strategy or quietly defeat it. The tool is only as useful as the ground it lands on, which is a lesson leaders already know from why most organisations fail at AI adoption: capability without the conditions to absorb it delivers little.

How do I capture the value without spraying AI everywhere?

With restraint and sequence, because the review's evidence rewards focus rather than breadth.

1. Start where the data is deepest and the decision repeats. Forecasting, routing and risk before anything else. That is where the measured returns cluster.
2. Embed AI into the process, not beside it. The review's guidance is to build AI into your business processes and run analytics close to where the data is created, so decisions happen in real time rather than in a monthly report.
3. Build absorptive capacity. Train people broadly, including those outside IT, to understand what AI can and cannot do. A workforce that understands the tool is what lets the value land, and it becomes a source of designing the loops the work runs in rather than merely operating them.
4. Keep a human check on the consequential calls. Automate the volume; hold judgement on the decisions that carry real risk across a border or a culture.
5. Choose by decision value, not tool novelty. A capability that sharpens a decision you make often earns its place; one that merely impresses does not.

What does this change for me as a leader this quarter?

It turns a vague pressure to "do something with AI" into a precise question: which two or three operational decisions, made often and on good data, would improve most if AI sharpened them? For most global businesses the honest answer starts in the supply chain and the market-selection process, because that is where the review finds the evidence strongest and the returns most measurable.

The competitors who look prescient are rarely doing something exotic. They have pointed a capable tool at a decision that matters, embedded it into how the work actually flows, and trained their people to trust and question it in equal measure. That is a sequence any leader can follow, and the best quarter to start following it is the one you are in.

About the author

Thomas Green

British technology futurist, AI keynote speaker and advisor. Thirty years across enterprise technology and AI strategy, helping leaders navigate the future of work. The futurist who died.