AI adoption: do success stories distort the picture in Basel?
When reporting progress on AI adoption, a manager will readily select a few convincing examples. The more demanding management question is how to prevent that selection from turning a collection of successes into a misleading assessment.
What the literature establishes
Jerker Denrell (2003, Organization Science) shows, through theoretical analysis and simulations, that the underrepresentation of failures can distort learning from others’ experience. Certain practices may appear associated with success because the observable cases form a selected sample. For AI, this suggests distinguishing the examples available for demonstration from the full set of attempts on which an assessment should draw.
Learning when experience is scarce
James G. March, Lee S. Sproull and Michal Tamuz (1991, Organization Science) analyse how organisations learn when observable events are rare. They examine, among other possibilities, how experience can be enriched through multiple interpretations and consideration of what might have happened. Applied to an AI trial, this reasoning encourages attention to revisions and discarded alternatives rather than only the final deliverable.
The premature conclusion
One might conclude that adding a few failures to presentations of success is enough. Yet a balanced collection of anecdotes remains a selection: it does not reveal how frequently outcomes occur across all attempts. The point is therefore not to discourage demonstrations, but to avoid treating them as a representative assessment. (our executive and employee training programmes)
What these studies cannot establish
These articles concern neither generative AI nor Basel-based companies; they illuminate mechanisms of organisational learning without measuring their magnitude in this setting. Nor is an abandoned attempt necessarily a failure of the tool: it may reflect a poorly defined request or a prudent decision to stop. A useful record must therefore distinguish the observed outcome from the reason attributed to it.
A practical check in Bâle
As part of SHR — Swiss Human Resources’ programme « Conduire l’adoption de l’IA dans son équipe », a Basel-based manager can define an observation period and an authorised task: for example, drafting a non-sensitive coordination note involving a pharmaceutical or life sciences team, a Rhine logistics partner and an international R&D unit. A minimal register records every attempt within that scope, including interrupted attempts, and distinguishes outputs accepted as they stand, revised or discarded, without recording confidential content. The measure to check is the proportion of outputs accepted without revision across all attempts, rather than only the examples presented. The manager can then compare this proportion with that of the cases selected for the report: a difference makes the selection effect visible, without by itself establishing the tool’s value. To go further: explore the Leading AI Adoption in Your Team training in Basel in the canton of Basel-Stadt, or browse our executive and employee training programmes in Switzerland.
In pictures: Leading AI Adoption in Your Team in Basel
- leading ai adoption in your team
- Basel
- research
- training basel