Operational Trends | #04: Buying AI Is Easy, Cleaning the Data Is Hard.

Welcome to the unique weekend article for the Loyal Fan subscribers-only edition.

You Hire the Best Worker, Then Seat Them Where They Cannot Work

Picture this. You have just hired the best craftsman in the region, someone who can do what no one else can. But instead of giving them a proper place to work, you sit them down at a bench piled high with junk. Tools lie jumbled together, the one you need is nowhere in sight, the ones you do not need are heaped up. Half the tools are broken and no one has marked them. Everyone before has left things their own way, so nothing has a fixed place. That brilliant worker, however gifted, has to spend most of their time digging, cleaning, and guessing what still works. Then you stand back, disappointed, and conclude that this worker turned out to be rather ordinary after all. But the problem was never the worker. The problem is the bench.

This is exactly what is happening with artificial intelligence in operations, and the broad numbers from the past week lay it bare to a startling degree. The PwC Digital Trends in Operations 2026 survey, conducted among 767 operations and supply chain leaders at US companies with revenue of one hundred million dollars or more, reveals a paradox so neat it reads like a joke: 85% consider themselves ahead of most competitors on digital transformation, while 89% admit their technology investments have not delivered the results they expected. Only 27% say they have embedded an AI strategy across the whole enterprise. And when asked what is blocking the value, the answer is not that the AI model is not smart enough, it is the complexity of integration, followed by data quality. In other words, a whole industry has hired the best worker, then sat it down at a cluttered bench, and is now disappointed that it is slow.

One thing must be said right away, to keep to the standard of honesty and to make this a piece about a trend rather than a feeling. One company complaining is only one complaint, but when many broad surveys all point the same way, that is a real trend. Alongside PwC, the KPMG US supply chain 2026 survey of 462 leaders at companies with revenue of one billion dollars or more found that 78% plan to reach a moderate or greater level of operational autonomy by 2027, meaning the AI ambition is very large. And the MIT supply chain sustainability report published in mid October reveals an operational detail worth pondering: among automated warehouses that use AI, 78% report higher electricity use, a sign that automation poured onto a foundation that is not yet tidy may even amplify waste. The three sources together paint the same picture: AI ambition is at its peak, but the results are stuck, and the place they are stuck is not the intelligence of the machine, but the foundation it stands on. This piece will use a single operations tool, and a surprisingly humble one at that, to dissect that foundation and show how to clean it. That tool is 5S.