Citation :

Michael Omondi, Esther Jepchirchir, Catherine Chepkirui, Monica Awino, 2026. "Large Language Model–Driven Economic Narrative Intelligence for Automated Financial Knowledge Synthesis" ESP International Journal of Science, Humanities & Management Studies(ESP-IJSHMS)  Volume 4, Issue 3: 29-41.

Abstract :

The rapid advancement of Large Language Models (LLMs) has significantly enhanced the ability of computational systems to interpret, synthesize, and generate information from large volumes of heterogeneous financial and economic text. This paper introduces the preliminary concept of Large Language Model–Driven Economic Narrative Intelligence (LLM-ENI), a framework designed to automate the transformation of financial reports, central bank communications, regulatory disclosures, market news, analyst commentary, and macroeconomic narratives into structured and interpretable financial knowledge. Unlike conventional approaches that primarily rely on sentiment analysis or document summarization, the proposed framework emphasizes cross-source contextualization, temporal reasoning, evidence linking, contradiction detection, source reliability, and confidence-aware narrative generation. The framework integrates retrieval-augmented generation (RAG), domain-adapted financial language models such as FinBERT, financial knowledge graphs, event extraction, and quantitative indicators to generate evidence-grounded economic narratives. The study also examines critical challenges associated with LLM-based financial intelligence, including hallucination, source reliability, model bias, outdated information, numerical reasoning limitations, privacy, regulatory compliance, and accountability. LLM-ENI is conceptualized as an analytical layer between raw financial information and human decision-making, enabling the rapid synthesis of complex and continuously evolving information across financial markets and economic systems. By integrating textual, numerical, and graph-structured information while maintaining evidence provenance, the proposed approach has the potential to improve information accessibility, traceability, and analytical transparency. The paper concludes by identifying future research directions for developing reliable, scalable, graph-aware, and real-time financial knowledge-synthesis systems.

References :

[1] Abaci, D. (2019) ‘Fin BERT: Financial sentiment analysis with pre-trained language models’, arrive preprint, arXiv:1908.10063.

[2] Brown, T.B., Mann, B., Ryder, N., Cubbish, M., Kaplan, J.D., Dhariwal, P., Neelakantan, A., Sham, P., Satyr, G., Gaskell, A. and Agawam, S. (2020) ‘Language models are few-shot learners’, Advances in Neural Information Processing Systems, 33, pp. 1877–1901.

[3] Devlin, J., Chang, M.W., Lee, K. and Tout nova, K. (2019) ‘BERT: Pre-training of deep bidirectional transformers for language understanding’, Proceedings of NAACL-HLT, pp. 4171–4186.

[4] Goo, H., Yu, H., Liu, Z., Li, X. and Zhang, Y. (2024) ‘Fight: Open-source financial large language models’, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pp. 1234–1246.

[5] Lewis, P., Perez, E., Pintos, A., Petron, F., Capuchin, V., Goal, N., Kittler, H., Lewis, M., Yam, W.T., Rocktäschel, T., Riedel, S. and Kiel, D. (2020) ‘Retrieval-augmented generation for knowledge-intensive NLP tasks’, Advances in Neural Information Processing Systems, 33, pp. 9459–9474.

[6] Lopez-Lira, A. and Tang, Y. (2023) ‘Can Catgut forecast stock price movements? Return predictability and large language models’, arrive preprint, arXiv:2304.07619.

[7] Ploughman, T. and McDonald, B. (2011) ‘When is a liability not a liability? Textual analysis, dictionaries, and 10-Ks’, The Journal of Finance, 66(1), pp. 35–65.

[8] Ploughman, T. and McDonald, B. (2016) ‘Textual analysis in accounting and finance: A survey’, Journal of Accounting Research, 54(4), pp. 1187–1230.

[9] Manning, C.D., Raghavan, P. and Schultz, H. (2008) Introduction to Information Retrieval. Cambridge: Cambridge University Press.

[10] Open-air (2023) ‘GPT-4 technical report’, arrive preprint, arXiv:2303.08774.

[11] Pang, B., Zhu, Y., Cui, J., Fu, H., Li, C., Wu, Y. and He, L. (2024) ‘Graph retrieval-augmented generation: A survey’, arrive preprint, arXiv:2408.08921.

[12] Qin, Y., Liang, S., Ye, X., Zhu, X., Yan, Y., Li, X., Tao, X., Tang, Y., Zhang, Y., Huang, Y., Deng, J., Zhang, Y., Chen, W., Liu, H., Sun, M. and Qin, B. (2024) ‘Tool learning with foundation models’, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pp. 1–20.

[13] Raffle, C., Shaker, N., Roberts, A., Lee, K., Nearing, S., Matera, M., Zhou, Y., Li, W. and Liu, P.J. (2020) ‘Exploring the limits of transfer learning with a unified text-to-text transformer’, Journal of Machine Learning Research, 21(140), pp. 1–67.

[14] Rahul, T., Nair, S., Khalid, A., Elavarasan, R.M., Raghavan, V., Thakur, R. and Kauri, P. (2023) ‘The impact of Catgut on business and finance: Opportunities and challenges’, Journal of Business Research, 166, 114079.

[15] Sheathe, Y.R., Ben-Mena hem, S.M. and von Krogh, G. (2019) ‘Organizational decision-making structures in the age of artificial intelligence’, California Management Review, 61(4), pp. 66–83.

[16] Fetlock, P.C. (2007) ‘Giving content to investor sentiment: The role of media in the stock market’, The Journal of Finance, 62(3), pp. 1139–1168.

[17] Aswan, A., Shaker, N., Parma, N., Uszkoreit, J., Jones, L., Gomez, A.N., Kaiser, Ł. and Polosukhin, I. (2017) ‘Attention is all you need’, Advances in Neural Information Processing Systems, 30, pp. 5998–6008.

[18] Wu, S., Irony, O., Lu, S., Dabravolski, V., Dredge, M., German, S., Kambadur, P., Rosenberg, D. and Mann, G. (2023) ‘BloombergGPT: A large language model for finance’, arrive preprint, arXiv:2303.17564.

[19] Yang, H., Liu, X.Y. and Wang, C.D. (2023) ‘Fight: Democratizing internet-scale data for financial large language models’, arrive preprint, arXiv:2307.10485.

[20] Zhang, J., Zhang, J., Cu, Q. and Chen, X. (2024) ‘Large language models in finance: A survey’, ACM Computing Surveys, 57(5), pp. 1-38.

Keywords :

Large Language Models, Economic Narrative Intelligence, Financial Intelligence, Financial Natural Language Processing, Retrieval-Augmented Generation, Financial Knowledge Graphs, Generative AI, Knowledge Synthesis, Narrative Analysis, Financial Information Extraction .