Volume 4,Issue 3
Research on the Collaborative Mechanism and Effectiveness Evaluation of AI Agents in End-to-End Enrollment Operations
Against the backdrop of the sustained advancement of education digitalization, higher education enrollment has placed higher demands on precision and efficiency. Because traditional enrollment models are fragmented across stages and isolated in terms of data, they can no longer meet the current needs of student-source competition and talent selection. Based on the realities of enrollment practice, this study identifies the types of AI agents suitable for the entire enrollment process, clarifies their functional positioning at each stage, further constructs a collaborative operating mechanism coveringpromotion, consultation, conversion, management, and other stages, and conducts analysis in combination with the practices of multiple institutions. The findings show that this model can connect the entire enrollment process, improve the accuracy of student-source identification, accelerate consultation feedback, enhance the scientific nature of enrollment decision-making, effectively control operating costs, and reduce operational risks. It provides a clear approach for the digital transformation of enrollment and offers practical methodology for improving enrollment quality and optimizing management.
[1] Chen XM, Hu S, Qi RM, 2025, International References and Local Adaptation of the “Admission - Education - Employment” Linkage Mechanism in Higher Education. Education and Career, (22): 67-74.
[2] Du MY, 2024, Research on Optimization of College Admission Promotion Strategies from the Perspective of Digital Transformation. Journal of Jilin Radio and Television University, (3): 128-130.
[3] Wang HQ, Cao Q, 2025, Admission Data Governance in Higher Vocational Colleges: Key Issues, Multiple Logics and Promotion Strategies. Big Data Era, (9): 43-49.
[4] Mensah AD, Ghebreziabiher KA, Boateng OG, et al., 2026, AI-native cloud-edge orchestration for 6G metaverse networks: an LLM-guided multi-agent DRL approach. Complex & Intelligent Systems, 12(4): 141.
[5] Han J, Othman BJ, Kim H, 2026, RIS-augmented distributed crowdsourcing multi-agent reinforcement learning with edge AI-enabled UAV communications. Journal of Systems Architecture, 175: 103727.
[6] Simpson J, Patil G, Stening H, et al., 2026, Can an AI agent lead human teams? Computers in Human Behavior: Artificial Humans, 7: 100278.
[7] Radanliev P, Santos O, Maple C, 2026, Threats and vulnerabilities in artificial intelligence and agentic AI models. Frontiers in Artificial Intelligence, 9: 1731566.