SmartSlide AI: An Intelligent Chatbot-Driven System for Automated PowerPoint and PDF Presentation Generation

This paper presented SmartSlide AI, a conversational chatbot-driven system that turns a natural-language request into an editable, dual-exported (PPTX + PDF) slide presentation. By staging generation into requirement extraction, editable outlining, layout-aware content generation, and a strictly native-object export engine, the system produces genuinely editable office files rather than rendered images, while a deterministic Demo AI Mode and defensive JSON-repair logic keep every code path demonstrable and crash-free even without a live LLM subscription. An automated test suite (16/16 passing), zero OOXML validation errors, and sub-two-second file-generation latency across the tested range of slide counts support the system's readiness as a working MCA-level prototype, and the identified future extensions — particularly retrieval-augmented, citation-grounded content generation — offer a clear path toward a more research-oriented deployment.

[1]  A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention Is All You Need,” in Proc. 31st Conf. Neural Information Processing Systems (NeurIPS), 2017. [2] T. Brown, B. Mann, N. Ryder, et al., “Language Models are Few-Shot Learners,” in Proc. 34th Conf. Neural Information Processing Systems (NeurIPS), 2020. [3] J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,” in Proc. NAACL-HLT, 2019. [4]   P. Lewis, E. Perez, A. Piktus, et al., “Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks,” in Proc. 34th Conf. Neural Information Processing Systems (NeurIPS), 2020. [5]  OpenAI, “GPT-4 Technical Report,” arXiv:2303.08774, 2023. [6]   G. Team, Google DeepMind, “Gemini: A Family of Highly Capable Multimodal Models,” arXiv:2312.11805, 2023. [7]    Anthropic, “Model Cards and Documentation,” Anthropic PBC, 2024. [Online]. Available: https://www.anthropic.com [8]     S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural Computation, vol. 9, no. 8, pp. 1735–1780, 1997. [9]   I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. Cambridge, MA, USA: MIT Press, 2016. [10]     Ecma International, “Office Open XML File Formats,” ECMA-376, 5th ed., 2021.