The Finding That Deserves More Attention

Business Insider recently asked eight supply chain leaders what impressed and surprised them most about AI deployment in their operations. The answers were revealing, but not in the way the headline suggests. Leaders like J-Ann Tio Toles described genuine operational transformation in logistics and sustainability tracking. What struck me reading through their responses was not the enthusiasm. It was the consistent pattern underneath it: every leader described being surprised by what AI could do. That surprise is the data point worth analyzing.

When experienced professionals with decades of supply chain expertise are genuinely surprised by the capabilities of tools their organizations have already deployed, something structural is happening. This is not a story about AI being impressive. It is a story about a persistent gap between awareness and capability that I have been tracking in a different domain, and which appears to generalize further than platform research has previously acknowledged.

Surprise as Diagnostic Evidence

Algorithmic literacy research has consistently documented what Kellogg, Valentine, and Christin (2020) describe as the layered opacity of algorithmic systems at work. Workers develop awareness that automated systems govern outcomes without developing accurate models of how those systems actually function. The supply chain leaders in this report confirm that pattern, but they add an important dimension. These are not entry-level workers. They are senior decision-makers with direct budget authority over the AI systems that surprised them.

This matters for organizational theory because it challenges a common assumption: that hierarchical position correlates with schema accuracy. The classical coordination literature treats expertise as roughly proportional to organizational rank, at least within a domain. What the supply chain evidence suggests is that algorithmic systems create what I would call a schema displacement effect. The structural features of AI-mediated logistics do not map cleanly onto the structural features of pre-algorithmic supply chain management. Experience in the old topology does not transfer automatically to the new one.

Gentner's (1983) structure-mapping theory is useful here. Transfer occurs when learners identify relational correspondences between a source domain and a target domain, not when surface features look similar. Supply chain expertise built around human-negotiated vendor relationships, manual inventory forecasting, and sequential decision loops shares almost no relational structure with AI-mediated systems that optimize across thousands of variables simultaneously. The experienced leader is in the same epistemic position as the novice when the underlying relational structure has changed. Surprise is the predictable result.

The Distinction Between Adaptive and Routine Expertise

Hatano and Inagaki (1986) drew a foundational distinction between routine expertise, which is the capacity to execute known procedures reliably, and adaptive expertise, which is the capacity to apply principles flexibly in novel conditions. The supply chain leaders who reported surprise were almost certainly drawing on routine expertise built for a pre-algorithmic environment. Their procedural knowledge remained intact and valuable in certain respects. Their structural schemas did not transfer.

This distinction has direct implications for how organizations should approach AI deployment. The current dominant model is platform-specific procedural training: teach workers the interface, the dashboard, the workflow. This produces faster initial performance. The ALC framework I am developing predicts that it will also produce brittle performance under novel conditions, because procedural training does not build the relational schemas that enable transfer. When the AI system updates, or when a different AI platform is adopted, the procedurally trained worker is back at the beginning. The supply chain leaders who described genuine transformation were, in nearly every case, describing moments where they stopped trying to map AI behavior onto familiar procedures and started reasoning about the structural logic of what the system was actually optimizing.

What Organizations Are Getting Wrong in AI Rollouts

The supply chain report implicitly describes a deployment pattern that is widespread: organizations invest heavily in AI infrastructure and then train workers on tool-specific procedures rather than on the underlying coordination logic. Schor et al. (2020) documented similar dynamics in platform labor markets, where workers develop folk theories about algorithmic behavior rather than accurate structural understanding. The folk theory is the procedural response to opacity. It is also, as the evidence consistently shows, an inadequate one.

What the eight supply chain leaders described as "surprise" is more precisely described as the moment when a folk theory fails and something closer to structural understanding begins to form. That transition, from procedural response to relational schema, is not guaranteed by exposure alone. It requires deliberate schema induction. Organizational rollouts that skip this step are not deploying AI effectively. They are deploying AI and then waiting for their workers to stumble into adaptive expertise by accident.

The surprise in that Business Insider report is not evidence of AI's power. It is evidence of a coordination problem that organizations have not yet named clearly enough to solve.

References

Gentner, D. (1983). Structure-mapping: A theoretical framework for analogy. Cognitive Science, 7(2), 155-170.

Hatano, G., and Inagaki, K. (1986). Two courses of expertise. In H. Stevenson, H. Azuma, and K. Hakuta (Eds.), Child development and education in Japan (pp. 262-272). Freeman.

Kellogg, K. C., Valentine, M. A., and Christin, A. (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366-410.

Schor, J. B., Attwood-Charles, W., Cansoy, M., Ladegaard, I., and Wengronowitz, R. (2020). Dependence and precarity in the platform economy. Theory and Society, 49(5-6), 833-861.