Your supply chain is ready for AI when it passes six practical tests: consistent data, systems that talk to each other, one defined source of truth, a clear decision each use case will change, teams that will actually trust and act on the output, and a named owner for governance. Most organisations pass some of these and fail others; the value of checking all six explicitly is knowing exactly which AI use cases are ready to start now, and which need groundwork first.
“Is AI ready for us?” is the wrong question. The AI itself is mature enough for most of the use cases businesses actually want. The right question is “are we ready for AI?”, and that answer varies enormously by organisation, often within the same business from one team to the next. AI readiness is a specific, checkable state, not a vague aspiration, and testing for it honestly before committing budget is the single best predictor of whether an AI initiative delivers or quietly stalls after the pilot.
It’s tempting to start by picking the most exciting AI use case (predictive maintenance, dynamic slotting, demand sensing) and work backwards from there. In our experience this is exactly backwards. Two businesses can want the same use case and get completely different results, because one has consistent, trustworthy data behind it and the other doesn’t. An AI readiness checklist isn’t a compliance exercise; it’s how you find out, before spending the budget, whether a given use case will actually work in your environment.
These six criteria come from work across hundreds of supply chain and logistics engagements. None of them require a perfect score to proceed, but an honest answer to each tells you exactly where to start, and where to do groundwork first.
The AI-readiness checklist: Six Criteria to Test Before Investing in AI
Does the same entity (a SKU, a customer, a site) mean the same thing, with the same identifier, across every system that touches it? This is the single most common readiness gap. A SKU with three different codes across ERP, WMS and finance isn’t a minor inconvenience for AI; it silently breaks the joins most use cases depend on.
Can data move between your WMS, ERP, TMS and finance systems without manual re-entry or translation? If a report only exists because someone manually rebuilds it in a spreadsheet every week, that process is a readiness gap, not a workaround; it means the underlying systems aren’t actually talking to each other.
For every data point a use case will rely on, is there one system, or one common data model, that is authoritative, rather than two or three that quietly disagree? Businesses rarely notice this gap until an AI model surfaces it, because people have learned to informally decide which number to trust.
Will the output feed a specific decision someone actually makes (a replenishment quantity, a shift roster, a carrier selection) or will it become another dashboard nobody consistently acts on? AI readiness isn’t only technical; it includes being specific about what changes as a result.
Will the people expected to act on an AI-driven recommendation actually trust it enough to use it? A model that’s statistically correct but operationally ignored delivers zero value. This usually means involving the team who will use the output early, not just at go-live.
Is there a named owner responsible for monitoring model performance and data quality over time, not just at launch? Data drifts, systems change, and a model that was accurate at go-live can quietly degrade without someone accountable for watching it.
Most organisations pass two or three of these criteria comfortably and fail the rest, which is normal, not disqualifying. The point of an AI readiness checklist isn’t to declare a business “not ready” and stop there; it’s to identify precisely which use cases can start now against the criteria you already meet, and which need targeted groundwork first, usually on data quality and source-of-truth questions. We cover that groundwork, and the specific, recurring problem of SKU identity mismatches between ERP and WMS, in our AI in Supply Chain white paper, alongside a full breakdown of the practical use cases this readiness work unlocks.
Readiness is deliberately positioned ahead of individual use cases in our AI in Supply Chain series. If AI inventory management or AI labour analytics sound like the right starting point operationally, this checklist is the fastest way to sanity-check whether the data and systems behind them are actually ready, before scoping a pilot around them.
What does it mean for a supply chain to be “AI-ready”?
It means the organisation passes six practical tests: consistent data, interoperable systems, a defined source of truth, a clear decision the AI output will change, people who will trust and act on the output, and named governance. It’s a checkable state, not a general sense of technological maturity.
Do we need to pass all six readiness criteria before starting an AI project?
No. Very few organisations pass all six on the first assessment. The value of checking each one explicitly is knowing which use cases are ready to start now against the criteria you already meet, and which need targeted groundwork first, usually data consistency or source-of-truth issues.
What’s the biggest AI readiness gap we typically see?
Data quality and consistency, most often surfacing as the same product having different identifiers across ERP and WMS. It’s rarely visible in daily operations because people learn to work around it, but it silently breaks the data joins most AI use cases depend on.
How long does an AI readiness assessment take?
A focused assessment against these six criteria, covering the systems and data relevant to a specific use case, is typically a matter of weeks rather than months, deliberately scoped to unblock a decision on where to start, not to become a project in itself.
Written by SCCG’s Team, drawing on the firm’s 700+ supply chain and logistics projects across 50+ countries, including extensive digital transformation, warehouse optimisation and logistics networks work.
Want an honest read on where your organisation stands against these six criteria?