Julian Waters-Lynch
Journal article · Open access

Shadow user innovation

Governing covert generative-AI use for dynamic-capability renewal

Julian Waters-Lynch and Darcy W. Allen and Jason Potts and Chris Berg

Employees are already using consumer-grade generative AI at work, often without formal approval, creating value while keeping the practice out of managerial sight.

What the paper argues

The paper defines shadow user innovation as covert, employee-led innovation enabled by concealable technologies such as generative AI. It sits between user innovation and shadow IT: employees privately experiment with tools that can improve their work, but the organisation may not see the learning, risk, or capability being created.

The paper models when hidden experimentation becomes visible enough for firms to learn from it. It links acute work-process pain, concealability, trust, expected gains, and sanction risk to the question of whether employees disclose what they have built or keep it underground.

Core mechanism: Shadow AI use becomes organisational capability only when firms can surface, evaluate, and integrate hidden experimentation

Why it matters

The paper matters because AI adoption is not only a formal strategy problem. It is also an everyday work-practice problem. Organisations that only police covert use may miss the most valuable evidence of where generative AI is actually useful.

Related product

From AI governance to auditable systems

Variorum extends the same concern with visible disagreement, traceability, and accountable AI judgement into manuscript review infrastructure.

Explore Variorum