The six jobs AI actually does — forecasting, inventory, maintenance, routing, negotiation, visibility — the boring NLP that powers them, and the honest line between old machine learning and the new, unreliable kind.
August 2026 · Case 2 of 6
As you read — hold this question
The brochures promise a self-driving supply chain. Strip the jargon away — what are the actual jobs AI does here, and which one of them is genuinely new?
6jobs
Everything the vendors sell reduces to six narrow jobs. None is a genius; each does one thing well.
It is a lot less science-fiction than the pitch decks suggest. Under the hood, “AI in the supply chain” is a toolbox of specific, unglamorous jobs — plus one boring branch of AI that quietly makes the rest possible, and one genuinely new capability that can confidently make things up.
The toolbox
The textbook versionQuiz: The six jobs & NLP
Six jobs, one boring hero, one tidy stack
Name the six jobs: forecasting demand; optimising inventory (safety stock and reorder points that move themselves); predictive maintenance (spotting a machine about to fail before it fails); optimising routes and networks; automated negotiation; and visibility (one live view over the whole chain). Put them together and they quietly rewire how the whole thing runs.
The least glamorous, most important part is natural-language processing (NLP): it reads the mountain of messy, unstructured paperwork — invoices, bills of lading, customs forms — and turns it into clean data the models can use. It is boring plumbing, and it is the thing that makes every shiny dashboard possible.
IBM’s “trifecta”
Data + AI + cloud, bolted on top of the creaky systems you already own — one pane of glass over the whole chain. Their advice: start small, but start.
The seven benefits (IBM)
Lower cost · real-time decisions · fewer errors · smarter inventory · a better-run warehouse · sustainability · simulate before you commit. (We audit this list in Case 4.)
The distinction that trips people up
The rest of the story
Machine learning vs. the chatty, unreliable kind
Almost everything above is machine learning — supervised models and optimisation that have run quietly in the background for years. Powerful, but not new, and not magic. The genuinely new bit is generative AI — the copilots and drafting tools — and it is newer, narrower, and it can confidently make things up.
The other side — the confident wrong answer
The real risk of generative AI in operations is not a killer robot; it’s a confident, wrong answer delivered with total certainty inside a decision that matters. Trust it like an eager intern whose work you check — not like an oracle. Brilliant assistant; unreliable oracle. Knowing which tool is which (old reliable ML vs. new fallible genAI) is the single most exam-useful distinction in the module.
Carry this
Name a few of the six jobs, say what NLP does (messy documents → clean shared data), and tell ordinary ML forecasting apart from a generative-AI copilot.
Quick recall — without looking back
Test yourself on this case
Question 1 of 3
Name at least four of the six AI jobs in the supply-chain toolbox.
Forecasting demand; optimising inventory (safety stock / reorder points); predictive maintenance; optimising routes & networks; automated negotiation; visibility (one live view of the chain).
Question 2 of 3
What does NLP do in this context, and why is it called the “boring hero”?
NLP reads unstructured paperwork — invoices, bills of lading, customs forms — and turns it into clean, structured data the models can use. It’s unglamorous plumbing, but every dashboard depends on it.
Question 3 of 3
Most supply-chain “AI” is which kind of AI — and what is the specific risk of the genuinely new kind?
Most is machine learning (supervised models / optimisation) — powerful but not new. The new kind is generative AI, whose risk is a confident, wrong answer stated with total certainty. Treat it as an intern to check, not an oracle.
Module 8 · Long Form · What AI Actually Does in Supply Chains
Module 8 · Short · What's Real, What's Hype
Sources
Module content
BUSN9049 Module 8 — lecture deck and videos (the six AI jobs; NLP; IBM “Put AI to work in supply chains” trifecta & seven benefits; SAS 2019). Flinders University, 2026.
ML vs generative AI
Innovare analysis, drawn from the module’s own machine-learning framing and the generative-AI “hallucination” literature.
Innovare Study
Long-form video: What AI Actually Does in Supply Chains. 2026. youtu.be/HlD8PWRNDiM