Engineering Autonomous Multi-Agent Software Systems: Implementing Hybrid Architectures, Interaction Protocols, and Execution Loops
Keywords:
Software Engineering 3.0, Multi-agent systems implementation, Agent-to-Agent protocols, Hybrid software architectures, AI TRiSM, System integration, Autonomous execution loopsAbstract
The paper examines the engineering transition from static software paradigms to autonomous agentic architectures (Software Engineering 3.0). Instead of focusing on organizational theory, the analysis concentrates on the technical implementation of functional subject attributes in software agents: goal planning, causal reasoning, and standardized execution. Key architectural patterns for deploying distributed multi-agent systems (MAS) are synthesized, specifically focusing on the integration of generative models with Case-Based Reasoning for strictly typed decision reproducibility. The study details the technical requirements for Agent-to-Agent (A2A) communication protocols and the establishment of stable behavioral contracts between autonomous entities. Implementation challenges are addressed through the lens of data engineering, specifically context memory management and vector data integration within existing IT landscapes. Furthermore, the paper structures the technical aspects of Artificial Intelligence Trust, Risk, and Security Management (AI TRiSM), defining methods for behavioral control, immutable logging, and decision traceability in high-load environments. The efficiency analysis is reframed from general business ROI to specific engineering metrics, correlating system performance with the costs of development, integration, and computational maintenance.
References
[1]. Ozgur Yayinlari. (n.d.). Navigating the shift from generative AI to autonomous agents in business strategy. Retrieved from: https://www.ozguryayinlari.com/site/catalog/download/1143/5485/12099 (date accessed: October 1, 2025).
[2]. McKinsey & Company. (2025, November 5). The state of AI: Global Survey 2025. Retrieved from: https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai (date accessed: November 6, 2025).
[3]. Gartner. (2025, August 5). Gartner hype cycle identifies top AI innovations in 2025. Retrieved from: https://www.gartner.com/en/newsroom/press-releases/2025-08-05-gartner-hype-cycle-identifies-top-ai-innovations-in-2025 (date accessed: October 3, 2025).
[4]. Hassan, A. E., et al. (2025). Agentic software engineering: Foundational pillars and a research roadmap(arXiv:2509.06216). https://doi.org/10.48550/arXiv.2509.06216 (date accessed: October 4, 2025). Retrieved from: https://arxiv.org/html/2509.06216v2
[5]. Luo, J., et al. (2025). Large language model agent: A survey on methodology, applications and challenges(arXiv:2503.21460). https://doi.org/10.48550/arXiv.2503.21460 (date accessed: October 5, 2025). Retrieved from: https://arxiv.org/abs/2503.21460
[6]. Gartner. (2025, October 20). Gartner identifies the top strategic technology trends for 2026. Retrieved from: https://www.gartner.com/en/newsroom/press-releases/2025-10-20-gartner-identifies-the-top-strategic-technology-trends-for-2026 (date accessed: October 22, 2025).
[7]. Benaich, N., & Air Street Capital. (2025). State of AI Report 2025. Retrieved from: https://www.stateof.ai/ (date accessed: October 7, 2025).
[8]. The Verge. (2025). Figma’s big AI update takes on Adobe, WordPress, and Canva. Retrieved from: https://www.theverge.com/news/662678/figma-buzz-draw-make-sites-announcement-availability (date accessed: December 12, 2025).
[9]. Gartner. (n.d.). Multiagent systems in enterprise AI: Efficiency, innovation and vendor advantage. Retrieved from: https://www.gartner.com/en/articles/multiagent-systems (date accessed: October 9, 2025).
[10]. ICML. (2025). Multi-Agent Systems in the Era of Foundation Models: Opportunities, Challenges and Futures (ICML 2025 Workshop). Retrieved from: https://icml.cc/virtual/2025/workshop/39955 (date accessed: October 10, 2025).
[11]. Hu, J., Dong, Y., Ao, S., Li, Z., Wang, B., Singh, L., Cheng, G., Ramchurn, S. D., & Huang, X. (2025). Position: Towards a responsible LLM-empowered multi-agent systems (arXiv:2502.01714). https://doi.org/10.48550/arXiv.2502.01714(date accessed: October 11, 2025). Retrieved from: https://arxiv.org/pdf/2502.01714
[12]. IEEE ICDCS 2025. (2025). Tutorial: Distributed multi-agent AI systems: Scalability, challenges, and applications. Retrieved from: https://icdcs2025.icdcs.org/tutorial-distributed-multi-agent-ai-systems-scalability-challenges-and-applications/ (date accessed: October 12, 2025).
[13]. Krishnan, N. (2025). Advancing multi-agent systems through model context protocol: Architecture, implementation, and applications (arXiv:2504.21030). https://doi.org/10.48550/arXiv.2504.21030 (date accessed: October 13, 2025). Retrieved from: https://arxiv.org/html/2504.21030v1
[14]. IBM. (2025). AI agents in 2025: Expectations vs. reality. Retrieved from: https://www.ibm.com/think/insights/ai-agents-2025-expectations-vs-reality (date accessed: October 14, 2025).
[15]. Google for Developers Blog. (2025). Introducing A2UI: An open project for agent-driven interfaces. Retrieved from: https://developers.googleblog.com/en/introducing-a2ui-an-open-project-for-agent-driven-interfaces/ (date accessed: December 16, 2025).
[16]. Supabase. (n.d.). Architecture | Supabase Docs. Retrieved from: https://supabase.com/docs/guides/getting-started/architecture (date accessed: October 15, 2025).
[17]. Composio. (n.d.). How to integrate Supabase MCP with Vercel AI SDK. Retrieved from: https://composio.dev/toolkits/supabase/framework/ai-sdk (date accessed: October 16, 2025).
[18]. Google. (n.d.). A2UI (repository). Retrieved from: https://github.com/google/A2UI (date accessed: December 17, 2025).
[19]. TypeScript. (n.d.). TypeScript Handbook. Retrieved from: https://www.typescriptlang.org/docs/handbook/intro.html(date accessed: October 17, 2025).
[20]. Vercel. (2025, November 10). How to build AI agents with Vercel and the AI SDK. Retrieved from: https://vercel.com/kb/guide/how-to-build-ai-agents-with-vercel-and-the-ai-sdk (date accessed: November 11, 2025).
[21]. conf.researchr.org. (2025). AISM 2025 – AI for Software Modernization. Retrieved from: https://conf.researchr.org/home/ase-2025/aism-2025 (date accessed: October 18, 2025).
[22]. Rad, S., Darvish Rouhani, B., & Banaeian, H. (2025). A proposed hybrid conceptual model of artificial intelligence and enterprise architecture for digital transformation in banks: An approach to improving business processes. Annals of Process Engineering and Management, 2(4), 200–210. https://doi.org/10.48314/apem.vi.37 (date accessed: November 10, 2025). Retrieved from: https://apem.reapress.com/journal/article/download/37/79
[23]. Gartner. (2024, October 21). Gartner identifies the top 10 strategic technology trends for 2025. Retrieved from: https://www.gartner.com/en/newsroom/press-releases/2024-10-21-gartner-identifies-the-top-10-strategic-technology-trends-for-2025 (date accessed: October 23, 2025).
[24]. Li, Z., Wu, W., Guo, Y., Sun, J., & Han, Q.-L. (2025). Embodied multi-agent systems: A review. IEEE/CAA Journal of Automatica Sinica, 12(6), 1095–1116. https://doi.org/10.1109/JAS.2025.125552 (date accessed: October 19, 2025). Retrieved from: https://www.ieee-jas.net/article/doi/10.1109/JAS.2025.125552
[25]. Plaat, A., et al. (2025). Agentic large language models, a survey (arXiv:2503.23037). https://doi.org/10.48550/arXiv.2503.23037 (date accessed: October 20, 2025). Retrieved from: https://arxiv.org/abs/2503.23037
[26]. Tarasyuk, V., & Müller-Czygan, G. (2025). Making tacit knowledge available: Generative AI as a bridge between human expertise and explainable AI systems. Annals of Social Sciences & Management Studies, 12(2), 555835. https://doi.org/10.19080/ASM.2025.12.555835 (date accessed: November 12, 2025). Retrieved from: https://juniperpublishers.com/asm/pdf/ASM.MS.ID.555835.pdf
[27]. Deloitte. (2025, March 12). The state of generative AI in the enterprise. Retrieved from: https://www.deloitte.com/az/en/issues/generative-ai/state-of-generative-ai-in-enterprise.html (date accessed: November 13, 2025).
[28]. Weitzman, M., et al. (2024). Design principles for generative AI applications. In Proceedings of the CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3613904.3642466 (date accessed: October 24, 2025). Retrieved from: https://dl.acm.org/doi/10.1145/3613904.3642466
[29]. Akbar, M. A. (2025). Agentic AI in software engineering: Practitioner perspectives across the software development life cycle. SSRN. https://doi.org/10.2139/ssrn.5520159 (date accessed: October 25, 2025). Retrieved from: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5520159
[30]. Microsoft. (2025, November 25). Microsoft Agent 365 SDK for JavaScript overview. Retrieved from: https://learn.microsoft.com/en-us/javascript/api/agent365-sdk-node/agent365-overview?view=agent365-sdk-node-latest (date accessed: November 26, 2025).
[31]. Ant Design. (n.d.). Components overview. Retrieved from: https://ant.design/components/overview/ (date accessed: October 26, 2025).
[32]. Nielsen Norman Group. (2025, October 3). Aligning UX metrics with organizational goals. Retrieved from: https://www.nngroup.com/articles/ux-metrics-goals/ (date accessed: November 26, 2025).
[33]. Alccahuamani, L. A. P., et al. (2025). Hybrid web architecture with AI and mobile notifications to optimize incident management in the public sector (Preprints.org, 202512.1056). https://doi.org/10.20944/preprints202512.1056.v1 (date accessed: December 20, 2025). Retrieved from: https://www.preprints.org/manuscript/202512.1056
[34]. OpenAI. (2025, December 7). The state of enterprise AI | 2025 report [PDF]. Retrieved from: https://cdn.openai.com/pdf/7ef17d82-96bf-4dd1-9df2-228f7f377a29/the-state-of-enterprise-ai_2025-report.pdf (date accessed: December 10, 2025).
[35]. McKinsey Global Institute. (n.d.). AI: Work partnerships between people, agents, and robots. Retrieved from: https://www.mckinsey.com/mgi/our-research/agents-robots-and-us-skill-partnerships-in-the-age-of-ai (date accessed: October 27, 2025).
[36]. Reuters. (2025). Microsoft wants AI agents to work together and remember things. Retrieved from: https://www.reuters.com/business/microsoft-wants-ai-agents-work-together-remember-things-2025-05-19/ (date accessed: November 19, 2025).
[37]. Ahmed, A., & Imran, A. S. (2025). The role of large language models in UI/UX design: A systematic literature review(arXiv:2507.04469). https://doi.org/10.48550/arXiv.2507.04469 (date accessed: October 28, 2025). Retrieved from: https://arxiv.org/abs/2507.04469
[38]. Obsidian Security. (2025, October 23). The 2025 AI agent security landscape: Players, trends, and risks. Retrieved from: https://www.obsidiansecurity.com/blog/ai-agent-market-landscape (date accessed: October 29, 2025).
[39]. OWASP. (2025, December 9). OWASP Top 10 for agentic applications for 2026. Retrieved from: https://genai.owasp.org/resource/owasp-top-10-for-agentic-applications-for-2026/ (date accessed: December 11, 2025).
[40]. Microsoft. (2025, May 19). Microsoft Build 2025: The age of AI agents and building the open agentic web. Retrieved from: https://blogs.microsoft.com/blog/2025/05/19/microsoft-build-2025-the-age-of-ai-agents-and-building-the-open-agentic-web/ (date accessed: October 30, 2025).
[41]. National Institute of Standards and Technology. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0) [PDF]. Retrieved from: https://nvlpubs.nist.gov/nistpubs/ai/nist.ai.100-1.pdf (date accessed: November 20, 2025).
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Mykhailo Nykoliuk
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Authors who submit papers with this journal agree to the following terms.
