Year: 2026
Type: dissertation
Abstract: Enterprise Content Management (ECM) systems hold much of the unstructured content that organisations rely on in daily operations, including contracts, policies, records, reports, and correspondence. Despite long-term investment in ECM platforms, many organisations continue to face difficulties in finding documents, applying consistent metadata, monitoring compliance, supporting flexible workflows, and turning stored content into usable insight. As a result, ECM can become a source of operational friction and regulatory risk rather than a strategic information-management capability. This thesis examines how artificial intelligence can be integrated into ECM platforms to address these challenges without weakening governance, explainability, or compliance. It develops and evaluates a platform-neutral framework for AI-enhanced ECM, organised into five layers: Content Intelligence Core, Integration and Activation, Lifecycle Touchpoints, Governance and Ethics, and Organisational Enablement. The framework identifies the technical, lifecycle, governance, and organisational capabilities required to move AI use in ECM from isolated pilots toward sustained operation. The research follows a Design Science Research methodology. The framework was developed iteratively across five drafts, informed by literature review, practitioner input, and an Alfresco Enterprise Edition reference architecture. The reference architecture was used to examine how the framework’s integration patterns could be mapped onto a real ECM platform through events, APIs, metadata write-back, workflow integration, search, and audit mechanisms. Summative evaluation was conducted through semi-structured interviews with five practitioners from financial services, healthcare, public sector, manufacturing, and consulting. The findings were analysed through an established information systems success model. The evaluation indicates that the framework is relevant, coherent, and technically feasible for organisations with sufficient integration maturity. The evaluation also identifies important boundary conditions: model lifecycle management was raised as under-specified relative to the rest of the governance layer, and legacy-platform interoperability is limited because the framework assumes modern API and event surfaces. Operational cost modelling and live deployment validation are outside the scope of this conceptual study and are identified as priorities for further research. The thesis contributes a graduated human-in-the-loop governance model, a three-stage AI-ECM maturity roadmap with explicit gate criteria, and sector-specific guidance for adoption. Together, these outputs provide a structured basis for organisations seeking to adopt AI-driven content intelligence under realistic governance and operational constraints.
Source: Theseus (Ammattikorkeakoulujen)
Author: Jha, Rajeshkumar
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7163297831