Leveraging Structured Data Input for Effective Chatbot Integration in Enterprises

• Caio Siqueira PUC-Rio https://orcid.org/0009-0009-2808-3278
• Orlando Fonseca PUC-Rio https://orcid.org/0000-0002-0205-2483
• Giuliano Ferreira PUC-Rio https://orcid.org/0009-0004-7824-0191
• Omar Leiva PUC-Rio https://orcid.org/0000-0001-6067-8372

DOI: https://doi.org/10.5753/stil.2024.245343

Resumo

This paper introduces an approach for integrating structured data into chatbot applications. Utilizing our Mindmap tool, which hierarchically organizes data and maps nodes to actions, we developed an augmented JSON schema to improve chatbot contextual understanding and response accuracy. By applying the Langchain suite and Retrieval-Augmented Generation techniques, our method enhances data retrieval and processing from a vector store, significantly improving interaction relevance.

Palavras-chave: Chatbot Integration, Structured Data, RAG, Langchain

Referências

LangChain Documentation. Langchain: Building applications with llms through composability. n.d.

Lewis, P., Oguz, B., Rinott, R., Riedel, S., and Stenetorp, P. (2020). Retrieval-augmented generation for knowledge-intensive nlp tasks. In Proceedings of the 34th Conference on Neural Information Processing Systems (NeurIPS 2020), Vancouver, Canada.

OpenAI (2023). Gpt-4 technical report. Technical report, OpenAI.

Vercel, Inc. (2024). React Foundations: About React and Next.js. Next.js Documentation.

Weiying, K., Pham, D. N., Eftekharypour, Y., and Pheng, A. J. (2019). Benchmarking nlp toolkits for enterprise application. In PRICAI 2019: Trends in Artificial Intelligence: 16th Pacific Rim International Conference on Artificial Intelligence, Cuvu, Yanuca Island, Fiji, August 26-30, 2019, Proceedings, Part III 16, pages 289–294. Springer.

Wen, Y., Wang, Z., and Sun, J. (2023). Mindmap: Knowledge graph prompting sparks graph of thoughts in large language models. arXiv preprint arXiv:2308.09729.