LinkedIn·Tuesday, 11 August 2026·15d ago

"Don't AI stupid." That's how Lora Cecere, founder of Supply Chain Insights, describes what most companies are doing with AI in planning…

Horizon Solutions

3,905 followers

"Don't AI stupid." That's how Lora Cecere, founder of Supply Chain Insights, describes what most companies are doing with AI in planning right now. Her point: bolting an agent onto your existing planning system to do a simple task, then calling it AI transformation, isn't enablement. It breaks the first time it hits a case nobody governed for. The real work is underneath, at the data layer. Most planning systems still can't talk to each other: transportation optimises for lowest cost, distribution planning optimises for order reliability, and there's no shared model connecting the two. Sales holds "ship-to" data. Supply chain runs on "ship-from" data. Without a semantic layer reconciling them, no agent on top of that mess gives you a real answer. A few other things she said that stuck: - 80% of the data around a supply chain (weather, reviews, supplier quality) goes unused, and using it is one of the fastest ways to cut demand latency - Small and mid-size companies often outperform enterprises here, not because of bigger data science teams, but because they're closer to the customer and carry less organisational politics - Stop measuring forecast error. Measure forecast value add: whether an input actually made the plan better Full conversation, on the Supply Chain Planning Podcast - https://lnkd.in/gxHaj-DM Ben Wim #supplychainplanning #dataquality #supplychainai