Implementing AI-Driven Workflow Automation in Corporate Secretarial Roles

Thesis Abstract

The rapid digitization of corporate governance processes has heightened the demand for efficient, accurate, and compliant secretarial operations, yet traditional, manually driven workflows remain error-prone and costly in multinational organizations. This study investigates the implementation of AI-driven workflow automation within corporate secretarial roles, addressing the problem of inconsistent task sequencing, delayed regulatory filings, and suboptimal document management that compromise governance efficacy and compliance. The aim is to design, implement, and evaluate an AI-enabled workflow platform that automates routine secretarial tasks—minutes drafting, board-pack generation, statutory filings, and contract lifecycle management—while ensuring auditability, accuracy, and regulatory conformance. Specific objectives include (1) mapping current secretarial workflows and identifying automation-ready tasks, (2) developing an AI-enabled orchestration model that integrates natural language processing, decision rules, and robotic process automation, (3) assessing impact on process durations, error rates, and compliance timeliness, (4) evaluating user acceptance and organizational readiness, and (5) formulating a framework for governance and risk management in AI-assisted secretarial practice. A mixed-methods research design is employed. The quantitative strand uses a quasi-experimental design with two comparable business units (n = 240 records per unit) over a nine-month period one unit adopts the AI-driven workflow automation (treatment), and the other maintains existing processes (control). Primary instruments include system-generated metrics (cycle time, error rate, filing timeliness, audit log completeness) and standardized user satisfaction surveys (n = 120 respondents across roles). The qualitative strand uses semi-structured interviews with 18 secretaries, 8 legal/compliance officers, and 6 board secretaries to capture perceived usefulness, trust, and risk considerations, complemented by diary studies of 4 secretarial teams. Data collection occurs in multinational manufacturing and financial services firms to ensure cross-sector generalizability. Validity and reliability procedures include triangulation of system logs with survey data, pilot testing of instruments (? ? 0.80 for scales), and member checking for interview transcripts. Analytical techniques comprise regression analyses to determine predictors of processing efficiency, mixed-effects models to account for nested data structures, and difference-in-differences estimation to isolate AI automation effects on performance indicators. Thematic analysis follows Braun and Clarke guidelines to interpret qualitative data, while an integrated framework combines diffusion of innovations and the Technology Acceptance Model (TAM) to interpret adoption dynamics. A process-mining approach is applied to the workflow logs to identify bottlenecks and reconfigure the AI orchestration rules. The study also develops a conceptual model linking AI versatility, governance controls, and perceived risk to measurable governance outcomes. Expected findings indicate a statistically significant reduction in cycle times (mean decrease of 28%), lower error rates (reduction from 3.8% to 1.2%), and improved filing timeliness (on-time filings rising from 82% to 96%) in the treatment unit compared with the control. User acceptance is anticipated to be positively associated with perceived usefulness and trust in AI explanations, moderated by task complexity. Qualitative insights are expected to reveal nuanced tensions between automation benefits and regulatory scrutiny, emphasizing the need for robust audit trails, explainable AI components, and clear governance policies. The study will identify critical success factors, including data quality, domain-specific NLP capabilities for legal drafting, and the alignment of AI workflows with existing board governance cycles. Contributions to knowledge include a) empirical evidence on performance gains and risk management implications of AI-driven secretarial automation in real-world corporate settings, b) an integrated framework combining TAM, diffusion theory, and governance risk management tailored to secretarial practice, and c) a practical blueprint for designing, implementing, and auditing AI-enabled workflows in secretarial operations, including data lineage, model governance, and compliance checks. The main conclusion is that AI-driven workflow automation can substantially enhance efficiency and accuracy in corporate secretarial roles when accompanied by rigorous governance, transparent decision-making, and continuous user engagement. Recommendations include establishing standardized data quality protocols, implementing explainable AI modules for drafting and decision support, formalizing auditability requirements, and adopting staged deployment with ongoing training and governance reviews to sustain long-term benefits.