Shadow User Innovation: Governing Covert Generative-AI Use for Dynamic-Capability Renewal
TITLE: Shadow User Innovation: Governing Covert Generative-AI Use for Dynamic-Capability Renewal
AUTHORS: Julian Waters-Lynch; Darcy W. E. Allen; Jason Potts; Chris Berg
YEAR: 2025
JOURNAL: Innovation: Organization & Management
DOI: https://doi.org/10.1080/14479338.2025.2519546
PUBLICATION_STATUS: Published online ahead of print
FIELDS: innovation economics; organisational theory; technology governance; generative AI adoption
KEYWORDS: shadow user innovation; shadow IT; generative AI; dynamic capabilities; organisational experimentation; user innovation; AI adoption in firms
Summary
This paper studies a new organisational phenomenon emerging with the diffusion of generative artificial intelligence tools: employees privately experimenting with AI systems at work without formal organisational approval.
The paper introduces the concept of Shadow User Innovation (SUI) to describe this behaviour. Shadow user innovation occurs when employees use unauthorised or unofficial technologies to experiment with improvements to their work processes while deliberately avoiding organisational visibility.
Generative AI tools such as ChatGPT, Claude, or image generation systems significantly increase the prevalence of this behaviour because they are easily accessible, inexpensive, require little technical expertise, and can be used privately through web interfaces or personal devices.
While such experimentation can generate valuable innovations that improve productivity, workflow design, and knowledge generation within organisations, these innovations often remain hidden from management. This creates a governance problem: organisations cannot systematically learn from or scale innovations that remain concealed.
The paper develops a theoretical framework explaining why shadow innovation emerges, how employees decide whether to reveal or conceal their innovations, and how organisations should design governance systems that allow them to capture the benefits of decentralised experimentation while managing associated risks.
Core Research Question
How should organisations govern covert experimentation with generative AI conducted by employees outside formal organisational approval?
The paper addresses three specific questions:
- Why do employees engage in covert experimentation with generative AI tools?
- Under what conditions do employees reveal or conceal the innovations they develop?
- What governance mechanisms allow organisations to benefit from employee experimentation while managing technological and organisational risks?
Concept: Shadow User Innovation
The concept of shadow user innovation combines insights from two research literatures.
User Innovation
User innovation research demonstrates that users frequently develop solutions to problems they encounter in practice. These user-developed innovations can become major drivers of technological change.
Shadow IT
Shadow IT refers to technologies adopted and used inside organisations without approval from central IT departments or organisational governance structures.
Shadow user innovation occurs when employees use unauthorised technologies not merely for convenience but to experiment with new solutions to workflow problems, thereby generating new knowledge and innovations.
Mechanism: How Shadow Innovation Emerges
The paper proposes that shadow user innovation emerges when two conditions are present.
1. Work Process Pain
Employees encounter inefficiencies or constraints in their existing work processes. These may include repetitive administrative tasks, slow internal software systems, or information bottlenecks that reduce productivity.
These problems create incentives for employees to search for alternative technological solutions.
2. Technological Concealability
The technology must be usable without detection by organisational monitoring systems.
Generative AI tools possess this property because they are accessible through external web services and require no integration with internal organisational infrastructure.
When employees face workflow problems and have access to easily concealable technologies, covert experimentation becomes likely.
Innovation Revelation
Once employees develop innovations through covert experimentation, they must decide whether to reveal those innovations to management.
This decision depends on several factors.
Expected benefits of disclosure
Employees may reveal innovations if they expect recognition, promotion, or organisational support.
Risk of sanctions
Employees may conceal innovations if using unauthorised technologies could lead to disciplinary action.
Organisational culture
Firms that reward experimentation are more likely to see innovations revealed.
Trust in management
Employees must believe that managers will respond constructively to experimentation rather than punish it.
If these conditions are not present, innovations remain hidden and organisations lose opportunities to learn from them.
Governance Responses
The paper identifies four organisational governance responses to shadow user innovation.
Repressive Governance
Strict prohibition of unauthorised technology use combined with monitoring and enforcement mechanisms. This reduces technological risk but discourages experimentation.
Permissive Governance
Management tolerates informal experimentation but does not actively integrate these innovations into organisational systems.
Integrative Governance
Organisations actively encourage employees to reveal innovations and provide mechanisms through which useful innovations can be adopted more broadly.
Structural Governance
Organisations create institutional mechanisms supporting experimentation, such as innovation sandboxes, experimentation programs, or formal reporting channels for employee-developed solutions.
The paper argues that integrative and structural approaches are most likely to convert decentralised experimentation into organisational capabilities.
Contribution
The paper contributes to several research literatures.
User innovation
It extends the literature by demonstrating how innovation can occur covertly inside organisations rather than only among external users.
Innovation governance
It analyses how governance structures influence whether employee-developed innovations are revealed or concealed.
Generative AI adoption
It highlights how the accessibility and concealability of generative AI technologies change the organisational dynamics of technological experimentation.
Dynamic capabilities
The paper shows how decentralised experimentation by employees can contribute to organisational learning and capability renewal.
Citation
Waters-Lynch, Julian; Allen, Darcy W. E.; Potts, Jason; Berg, Chris.
2025.
Shadow user innovation: governing covert generative-AI use for dynamic-capability renewal.
Innovation: Organization & Management.
DOI: https://doi.org/10.1080/14479338.2025.2519546
Source
Publisher page:
Darcy Allen 