Author(s): Prateek Batla

Abstract: The rapid emergence of AI commerce agents is reshaping enterprise workflows, from billing, pricing, and promotions to fraud detection and customer personalization. While these agents promise major gains in efficiency and scale, their deployment raises critical concerns about autonomy, trustworthiness, and security in mission-critical environments. Prior research addresses trust in AI agents, workflow automation, and zero trust architectures separately, but few studies integrate these dimensions within the specific context of enterprise commerce. This paper proposes a comprehensive framework for trustworthy autonomy in AI commerce agents, synthesizing recent advances in agentic AI, workflow orchestration, and enterprise security. The framework specifies layered mechanisms across perception, reasoning, trust and security, and action, underpinned by transparency, accountability, fairness, and resilience. It operationalizes graded autonomy with human-in-the-loop and human-on-the-loop controls, policy-aware guardrails, and auditable safety cases. Drawing on recent works from journals and peer-reviewed venues, we show how the framework applies to subscription billing, dynamic pricing, cross-border payments, order management, and fraud prevention. I outline evaluation dimensions across technical, organizational, and ethical criteria and provide a practical rubric for enterprise adoption. By bridging autonomy and trust in enterprise contexts, the study contributes both a conceptual foundation and a deployable architecture for AI commerce agents at scale.

DOI: 10.33545/27076571.2025.v6.i2c.196

Pages: 169-176 | Views: 295 | Downloads: 93

Download Full Article: Click Here