0 cumulative citations
View corpus contextTechnology alone does not modernise government: without explicit, integrated and revisable process knowledge, AI and other digital investments can entrench poor practices, erode tacit skills and create path‑dependent failures; successful transformation depends on sequencing, reversibility and organisational translation rather than algorithmic sophistication.
Digital transformation and knowledge-based process management in public services: barriers, enablers and sociotechnical dynamics
openalex
review_meta
low evidence
7/10 relevance
Summary only
summary available; pdf_status=paywall
DOI
Source
PDF
Public-sector digital and AI initiatives succeed or fail primarily because of organisational translation — the three dimensions of explicitation, knowledge integration and revisability determine whether technical capacity becomes administrative capacity, and failures arise from decoupling, frozen legacy processes, tacit‑knowledge erosion and institutional misfit.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
Governments invest continuously in digital transformation, yet a large share of initiatives fails to convert technological capacity into administrative capacity. The literature has established that the obstacles are seldom technical, but it has not specified the organisational mechanism through which the conversion happens or fails. The gap is theoretical and explanatory rather than merely empirical. The article asks how barriers, enablers and sociotechnical dynamics condition digital transformation when the phenomenon is examined through knowledge-based process management. An integrative review was conducted following the five stages of Whittemore and Knafl (2005), with three cycles of thematic coding over studies on digital transformation, process management, knowledge management and public sector modernisation. The findings organise barriers and enablers into five symmetrical domains and identify four recurrent dynamics: decoupling between strategy and enacted process, digitisation of the procedural legacy, erosion of tacit knowledge under automation, and asymmetry between technical and institutional integration. The theoretical contribution lies in defining knowledge-based process management as a three-dimensional construct, comprising explicitation, knowledge integration and revisability, and in showing that it operates as a translation mechanism between the technological and the institutional layers of public administration. Five propositions formalise the mechanism and account for why dynamic capabilities produce uneven effects on performance in government. As an implication, the effectiveness of digital transformation depends on the sequence in which irreversible commitments are made rather than on the sophistication of the technologies adopted.
Summary
Main Finding
The paper argues that the bottleneck in public-sector digital transformation is primarily organisational — specifically, the failure to translate technological capacity into administrative capacity through knowledge-based process management. It defines knowledge-based process management as a three-dimensional translation mechanism (explicitation, knowledge integration, revisability) that mediates between technological and institutional layers. Conversion succeeds or fails according to how this mechanism is structured and to the sequence of irreversible commitments governments make; sophisticated technology alone does not guarantee better outcomes.
Key Points
- Empirical puzzle: many government digital initiatives fail to improve administrative performance despite ongoing investments; obstacles are rarely purely technical.
- Conceptual contribution: introduces knowledge-based process management as a 3‑dimensional construct:
- Explicitation — the degree to which process knowledge is made explicit and codified.
- Knowledge integration — the coordination and combining of knowledge across roles, units and systems.
- Revisability — the ability to update, correct or roll back process definitions and enacted routines.
- Four recurrent sociotechnical dynamics that drive failure modes:
- Decoupling between strategy and enacted process (plans do not become lived routines).
- Digitisation of procedural legacy (automation freezes suboptimal legacy procedures).
- Erosion of tacit knowledge under automation (loss of skills and judgment when tasks are codified).
- Asymmetry between technical and institutional integration (systems interoperate technically but not institutionally).
- Barriers and enablers are organised into five symmetrical domains (roughly: strategic/organisational, processual, knowledge-related, technical/IT, and institutional/regulatory).
- The paper formulates five propositions that formalise how the three dimensions of knowledge-based process management operate as the translation mechanism and explain why dynamic capabilities (e.g., investments, skills) produce uneven performance effects.
- Policy-relevant implication: effectiveness depends more on the ordering and reversibility of commitments during transformation than on the intrinsic sophistication of adopted technologies.
Data & Methods
- Method: integrative literature review following Whittemore & Knafl (2005).
- Process: three cycles of thematic coding across literatures on digital transformation, process management, knowledge management and public sector modernisation.
- Outcome: synthesis that combines thematic findings and theoretical integration to generate propositions and a conceptual mechanism (knowledge-based process management).
Implications for AI Economics
- For public-sector AI adoption, returns depend critically on organizational knowledge practices, not just algorithmic performance. AI investments risk locking in poor practices if explicitation and revisability are weak.
- Sequencing and reversibility matter: irreversible procurement or technical integrations (e.g., embedding ML models into workflows or national platforms) can freeze legacy procedures and generate path-dependent losses; economists should account for option value of reversibility when valuing projects.
- Tacit-knowledge erosion: automation and opaque AI can displace tacit skills (judgment, discretion), reducing absorptive capacity and long-run productivity gains — evaluations should include human-capital and institutional-capital depreciation.
- Decoupling and institutional fit: technical interoperability does not imply institutional integration. Economic models of public-sector productivity gains from AI should include institutional coordination frictions and principal–agent misalignments.
- Heterogeneous effects of dynamic capabilities: differences in explicitation, integration and revisability explain why similar AI investments produce varying outcomes across agencies/jurisdictions. Empirical work should model these mediating variables rather than treating capabilities as homogeneous inputs.
- Policy and design recommendations (for economists advising policy):
- Prioritise mechanisms that increase explicitation (structured documentation, clear process models) and ensure knowledge integration (cross-unit governance, data-sharing protocols).
- Preserve revisability: favor modular, reversible deployments and staged pilots that allow rollback and iterative redesign.
- Protect and transfer tacit knowledge alongside automation (shadowing, training, gradual automation, human-in-loop designs).
- Sequence commitments to ensure institutional alignment before large-scale technical integration (governance, incentives, legal/regulatory fit).
- Use evaluation frameworks that capture option value, path-dependence and long-run institutional costs, not only short-term efficiency gains.
- Research directions for AI economics:
- Empirically test the paper’s five propositions with panel data or matched case studies across agencies/countries.
- Quantify the welfare costs of digitising legacy procedures and the value of reversibility.
- Model dynamic investment decisions where explicitation, integration and revisability evolve endogenously and shape returns to AI.
- Study how AI-specific characteristics (opacity, learning dynamics, data dependence) interact with the three dimensions of knowledge-based process management.
If you want, I can (a) map the paper’s five propositions into a formal economic model sketch, (b) propose empirical strategies and datasets to test them, or (c) produce a one‑page checklist for policymakers implementing AI in government that operationalises explicitation, integration and revisability. Which would be most useful?
Assessment
Paper Typereview_meta
Evidence Strengthlow — The paper is an integrative literature review and conceptual synthesis that develops propositions and a three‑dimensional mechanism; it does not present original causal empirical evidence or identification, so claims are theoretical and inferential rather than causally established.
Methods Rigormedium — The authors follow a recognised integrative review method (Whittemore & Knafl) and report three cycles of thematic coding across multiple literatures, which supports systematic synthesis and theory-building; however, the approach lacks described quantitative validation, pre-registered search protocols or causal identification strategies, and is vulnerable to selection and interpretive bias.
SampleAn integrative literature review synthesized across literatures on digital transformation, process management, knowledge management and public sector modernisation, using three cycles of thematic coding to develop propositions and the 'knowledge‑based process management' conceptual mechanism.
Themesorg_design adoption human_ai_collab productivity
GeneralizabilityConceptual synthesis rather than primary empirical analysis — propositions require empirical testing across contexts., Focused on public sector organisations; applicability to private firms or different institutional settings may be limited., Conclusions depend on the body of literature selected and interpreted; possible selection and publication biases in source materials., Does not quantify magnitudes (costs/benefits) or provide causal estimates, limiting direct application for policy cost–benefit calculations., May under-specify heterogeneity across jurisdictions (legal, political, procurement) that affects AI deployment outcomes.
Claims (9)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| The primary bottleneck in public-sector digital transformation is organisational rather than purely technological: governments often fail to translate technological capacity into administrative capacity. Organizational Efficiency | negative | Administrative performance from digital transformation |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Knowledge-based process management functions as a three-dimensional translation mechanism between technological and institutional layers. Organizational Efficiency | positive | Translation of technological capacity into institutional and administrative capacity |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The effectiveness of digital transformation depends on how knowledge-based process management is structured and on the sequence of irreversible commitments made by governments; sophisticated technology alone does not guarantee better outcomes. Organizational Efficiency | mixed | Administrative and organisational outcomes of digital transformation |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Digitising procedural legacy can automate and freeze suboptimal legacy procedures. Organizational Efficiency | negative | Efficiency and quality of enacted administrative processes |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Automation can erode tacit knowledge by reducing skills and judgment when tasks are codified. Skill Obsolescence | negative | Retention of tacit skills and professional judgment |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Technical interoperability between systems does not necessarily produce institutional integration across organisations, roles, or units. Organizational Efficiency | mixed | Institutional coordination and integration of administrative processes |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The paper proposes that explicitation, knowledge integration, and revisability mediate the effects of dynamic capabilities such as investments and skills, helping explain uneven performance across public-sector transformations. Organizational Efficiency | mixed | Variation in performance effects from organisational capabilities and digital investments |
Reading fidelity
high
Study strength
low
|
not reported
|
| The literature review finds that many government digital initiatives fail to improve administrative performance despite continuing investment. Organizational Efficiency | negative | Administrative performance following digital investment |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The paper argues that transformation effectiveness depends more on the ordering and reversibility of commitments than on the intrinsic sophistication of the adopted technologies. Organizational Efficiency | mixed | Effectiveness of public-sector digital transformation |
Reading fidelity
high
Study strength
low
|
not reported
|