Elamur Research⌕
☰
home / papers / 2026 / Building a Conversational AI Assistant for African Travel Services with LLMs and RAG

Building a Conversational AI Assistant for African Travel Services with LLMs and RAG

Paper · 2026 · Grace Kevine Ngoufo, Shamsuddeen Hassan Muhammad and Kevin Jeff Fogang Fokoa

Summary

Many African travel agencies receive more customer questions than their small teams can answer, and the chatbots they use are often rigid, rule-based, or missing entirely. This system-demonstration paper from AIMS Cameroon and Imperial College London replaces a step-by-step chatbot with an LLM assistant that customers reach through WhatsApp. The assistant first detects the user's intent. Small talk is answered by the LLM directly. Flight requests run through a clarification dialogue and then a call to the agency's booking APIs via the Model Context Protocol. Other service questions use a RAG pipeline over the company's own documents. Staff keep this knowledge base current through a password-protected web portal, where they can upload, delete, and monitor text and PDF files. The system was deployed with a real agency. The authors report fast replies, context kept across sessions, and smooth switching between English and French. However, the evidence consists of example dialogues, not quantitative evaluation. The work is a practical template for LLM customer service in African settings with limited resources.

Key points

  • Core stack: Gemini-2.5-flash is the LLM, accessed through its API. LangChain community memory stores each user's conversation history. Prompt engineering and few-shot examples steer the assistant's behaviour.
  • Retrieval and integrations: Documents are embedded with BGE-M3, which handles multiple languages, and stored in the open-source Qdrant vector database. Qdrant was chosen over Pinecone, Weaviate, and Milvus. A Meta Webhook connects the system to WhatsApp, and MCP links it to APIs for IATA codes and live flight availability.
  • Conversation handling: The example dialogues show the assistant:
    • managing a shift from one traveller to another (father to daughter);
    • correcting a one-way booking to a round trip;
    • asking the user to confirm details before searching flights;
    • converting prices from INR to XAF;
    • returning booking links.
  • Motivation: The authors cite a 41% AI adoption rate in African contexts (King et al., 2025). They argue that existing local chatbots are mostly template-driven.
  • Future work: Planned additions are voice interaction (speech-to-text and text-to-speech for WhatsApp voice notes) and African-language support, starting with Swahili and Hausa.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Imperial College LondonLondon · United Kingdom48 works · 12,181 citations🔬African Institute for Mathematical SciencesLimbe · Cameroon2 works · 81 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Dialogue systems · 624 works

Cite

Cite as Grace Kevine Ngoufo, Shamsuddeen Hassan Muhammad and Kevin Jeff Fogang Fokoa. 2026. Building a Conversational AI Assistant for African Travel Services with LLMs and RAG. Proceedings of the 7th Workshop on African Natural Language Processing (AfricaNLP 2026).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.africanlp-main.16

@inproceedings{ngoufo2026building,
  title = {{Building a Conversational AI Assistant for African Travel Services with LLMs and RAG}},
  author = {Grace Kevine Ngoufo, Shamsuddeen Hassan Muhammad and Kevin Jeff Fogang Fokoa},
  booktitle = {Proceedings of the 7th Workshop on African Natural Language Processing (AfricaNLP 2026)},
  year = {2026},
  doi = {10.18653/v1/2026.africanlp-main.16},
  url = {https://doi.org/10.18653/v1/2026.africanlp-main.16}
}

References

26 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. David Ifeoluwa Adelani, Graham Neubig, Sebastian Ruder, Shruti Rijhwani, Michael Beukman, Chester Palen-Michel, Constantine Lignos, Jesujoba Alabi, Shamsuddeen H Muhammad, Peter Nabende, and 1 others. 2022. Masakhaner 2.0: Africa-centric transfer learning for named entity recognition. In Proceedings…
  2. Fahd Azaroual. 2024. Artificial intelligence in africa: Challenges and opportunities. Policy Brief. PB, pages 23–24. find
  3. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, and 1 others. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901.
  4. Reine Marie Ndéla Marone and Moustapha Mbengue. 2025. Chatbots and artificial intelligence to support digital university libraries in africa: Opportunities and challenges. Digital Libraries Across Continents, pages 72–92. find
  5. Jianlv Chen, Shitao Xiao, Peitian Zhang, Kun Luo, Defu Lian, and Zheng Liu. 2024. Bge m3-embedding: Multi-lingual, multi-functionality, multi-granularity text embeddings through self-knowledge distillation. arXiv preprint arXiv:2401.XXXXX. find
  6. Joshua Maynez, Shashi Narayan, Bernd Bohnet, and Ryan McDonald. 2020. On faithfulness and factuality in abstractive summarization. arXiv preprint arXiv:2005.00661.
  7. Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, and 1 others. 1. others. 2022. palm: Scaling language modeling with pathways. arXiv preprint arXiv:2204.02311.
  8. LangChain Community. 2023. Langchain: Building applications with llms. https://docs.langchain.com/oss/python/langchain/. ↗
  9. Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yixin Dai, Jiawei Sun, Haofen Wang, and Haofen Wang. 2023. Retrieval-augmented generation for large language models: A survey. arXiv preprint arXiv:2312.10997, 2(1).
  10. Michael F McTear. 2002. Spoken dialogue technology: enabling the conversational user interface. ACM Computing Surveys (CSUR), 34(1):90–169. find
  11. Wilhelmina Nekoto, Vukosi Marivate, Tshinondiwa Matsila, Timi Fasubaa, Taiwo Fagbohungbe, Solomon Oluwole Akinola, Shamsuddeen Muhammad, Salomon Kabongo Kabenamualu, Salomey Osei, Freshia Sackey, and 1 others. 2020. Participatory research for low-resourced machine translation: A case study in africa…
  12. Keneth Ogueji and 1 others. 2021. Afriberta: African multilingual language model. In NeurIPS Workshop on AfricanNLP. find
  13. Postman. 2024. What is an api? a beginner’s guide to apis. https://www.postman.com/ what-is-an-api/. ↗
  14. Google DeepMind. 2024a. Gemini 2 flash model card. https://modelcards.withgoogle.com/ assets/documents/gemini-2-flash.pdf. ↗
  15. Google DeepMind. 2024b. Gemini flash. https://deepmind.google/technologies/ gemini/flash/. ↗
  16. Gautier Izacard and Edouard Grave. 2021. Leveraging passage retrieval with generative models for open domain question answering. In Proceedings of the 16th conference of the european chapter of the association for computational linguistics: main volume, pages 874–880.
  17. Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, and 1 others. 2024. Mixtral of experts. arXiv preprint arXiv:2401.04088. find
  18. Stephen King, Toni Morrison, Ernest Hemingway, F Scott Fitzgerald, and Mark Twain. 2025. Digital innovation and automation in african enterprises. find
  19. Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, and 1 others. 2020. Retrieval-augmented generation for knowledge-intensive nlp tasks. Advances in neural information processing systems, 33:9459–…
  20. Y Li, J Chen, and X Wang. 2023. Prompt engineering for large language models: A survey. arXiv preprint.
  21. Model Context Protocol. 2023. What is the model context protocol (mcp)? https://modelcontextprotocol.io/docs/ getting-started/intro. ↗
  22. Qdrant Team. 2023. Qdrant: High-performance vector database for scalable ai applications. https://qdrant.tech. ↗
  23. Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, and 1 others. 2023. Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
  24. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems, 30.
  25. Joseph Weizenbaum. 1966. Eliza—a computer program for the study of natural language communication between man and machine. Communications of the ACM, 9(1):36–45. find
  26. Jun Xiao, Sheng Zhang, Zhuosheng Zhang, Fei Huang, and Heng Ji. 2024. Bge-m3: Multi-functionality, multi-linguality, and multi-granularity embedding model. https://bge-model.com/bge/ bge_m3.html. ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ngoufo-2026-building-conversational-ai-assistant-african/