Adaptive API Support: A Human-in-the-Loop Agentic RAG Framework for Enterprise Financial Systems
Keywords:
Retrieval-Augmented Generation, Human-in-the-Loop Validation, Enterprise Collaboration Platforms, Financial Technology APIs, Dual-Feedback LearningAbstract
The world of enterprise developer support in the context of financial technology is characterized by a critical situation where the artificial intelligence usage rate is increasing, both by speeding up response generation and, at the same time, creating a lack of trust that requires verification by a person before production is put into place. The Agentic Knowledge Orchestrator resolves this paradox by offering a Retrieval-Augmented Generation framework, which conceptually incorporates domain expert validation as an inherent part and not as an exception-handling system. The system enhances AI output by integrating Human-in-the-Loop approval gates directly into enterprise collaboration systems, which convert unverified AI outputs into knowledge artifacts that are validated by subject matter experts before being shared with developer communities. The framework itself works on two-feedback learning strategies that detect expert corrections in validation stages and track patterns of end-user acceptance after the deployment, and will refine the knowledge base on the basis of the authoritative domain knowledge and practical utility indicators. Architectural constraints based on empirical foundations, based on benchmark evaluations, guide reranking strategies, response synthesis budgets that are optimized to financial services settings where accuracy, auditability, and compliance governance are the key design parameters. The architecture seals the recorded divide between recent adoption of AI tools and an ongoing lack of trust in the automated systems by developers by ensuring that automated support systems are tied to human expertise, and yet provide the scalability required of a global enterprise. Such a symbiotic combination of machine smarts and human intuition creates a precedent that can be repeated in areas that are knowledge-intensive and in which error is unavailable, but scale necessitates automation, especially in financial technology ecosystems in which API advice has a role in shaping security-critical deployments, regulatory and compliance adherence, and transaction-processing integrity.
DOI: https://doi.org/10.17762/ijisae.v14i1s.8319
Downloads
References
Satya Nadella, “Microsoft Fiscal Year 2024 First Quarter Earnings Conference Call,” Microsoft, 2023. [Online]. Available: https://www.microsoft.com/en-us/investor/events/fy-2024/earnings-fy-2024-q1
Stack Overflow, “2024 Developer Survey – AI”. [Online]. Available: https://survey.stackoverflow.co/2024/ai
Stack Overflow, “2024 Developer Survey,” 2024. [Online]. Available: https://survey.stackoverflow.co/2024/
Kevinbrowne, "Verification debt is the AI era’s technical debt," 2025. [Online]. Available: https://www.kevinbrowne.ca/verification-debt-is-the-ai-eras-technical-debt/
Patrick Lewis et al., “Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks,” NeurIPS, 2020. [Online]. Available: https://proceedings.neurips.cc/paper_files/paper/2020/file/6b493230205f780e1bc26945df7481e5-Paper.pdf
Yue Wang et al., “Laboratory for Analytic Sciences in TREC 2024 Retrieval Augmented Generation Track,” NIST. [Online]. Available: https://trec.nist.gov/pubs/trec33/papers/ncsu-las.rag.pdf
Postman, “2024 State of the API Report,” 2024. [Online]. Available: https://voyager.postman.com/doc/postman-state-of-the-api-report-2024.pdf
Long Ouyang et al., “Training Language Models to Follow Instructions with Human Feedback,” arXiv:2203.02155, 2022. [Online]. Available: https://arxiv.org/abs/2203.02155
Xiao Yang et al., “CRAG – Comprehensive RAG Benchmark,” 38th Conference on Neural Information Processing Systems (NeurIPS 2024) Track on Datasets and Benchmarks, 2024. [Online]. Available: https://proceedings.neurips.cc/paper_files/paper/2024/file/1435d2d0fca85a84d83ddcb754f58c29-Paper-Datasets_and_Benchmarks_Track.pdf
Dora, “DORA State of AI-Assisted Software Development,” 2025. [Online]. Available: https://dora.dev/research/shared/dora-report-2025/
Downloads
Published
How to Cite
Issue
Section
License

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.


