<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">dsait</journal-id><journal-title-group><journal-title xml:lang="ru">Цифровые решения и технологии искусственного интеллекта</journal-title><trans-title-group xml:lang="en"><trans-title>Digital Solutions and Artificial Intelligence Technologies</trans-title></trans-title-group></journal-title-group><issn pub-type="epub">3033-7097</issn><publisher><publisher-name>Финансовый университет при Правительстве Российской Федерации</publisher-name></publisher></journal-meta><article-meta><article-id custom-type="elpub" pub-id-type="custom">dsait-8</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>ТЕМА НОМЕРА: Искусственный интеллект и машинное обучение</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>COVER STORY: Artificial intelligence and machine learning</subject></subj-group></article-categories><title-group><article-title>Мультимодальный Telegram-бот на базе LLM -оркестратора: архитектура, экономика лимитов и влияние на пользовательский опыт</article-title><trans-title-group xml:lang="en"><trans-title>Multimodal Telegram Bot Based on LLM Orchestrator: Architecture, Economics of Limits and Impact on User Experience</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0006-7590-1315</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Зайцев</surname><given-names>Д. А.</given-names></name><name name-style="western" xml:lang="en"><surname>Zaitsev</surname><given-names>D. A.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Денис Андреевич Зайцев - студент факультета международных экономических отношений</p><p>Москва</p></bio><bio xml:lang="en"><p>Denis A. Zaitsev -Student of the Faculty of International Economic Relations</p><p>Moscow</p></bio><email xlink:type="simple">denis3849484@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0000-4113-2147</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Прудников</surname><given-names>А. В.</given-names></name><name name-style="western" xml:lang="en"><surname>Prudnikov</surname><given-names>A. V.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Андрей Валерьевич Прудников - студент факультета международных экономических отношений</p><p>Москва</p></bio><bio xml:lang="en"><p>Andrey V. Prudnikov - Student of the Faculty of International Economic Relations</p><p>Moscow</p></bio><email xlink:type="simple">apr9553@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0009-0006-9848-7495</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Хрипунова</surname><given-names>М. Б.</given-names></name><name name-style="western" xml:lang="en"><surname>Khripunova</surname><given-names>M. B.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Марина Борисовна Хрипунова - кандидат физико-математических наук, доцент, доцент кафедры математики и анализа данных</p><p>Москва</p></bio><bio xml:lang="en"><p>Marina B. Khripunova - Cand. Sci. (Phys. and Math.), Ass. Prof., Ass. Prof. of Department of Mathematics and Data Analysis</p><p>Moscow</p></bio><email xlink:type="simple">mbkhripunova@fa.ru</email><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0003-4380-3850</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Шмелева</surname><given-names>Л. А.</given-names></name><name name-style="western" xml:lang="en"><surname>Shmeleva</surname><given-names>L. A.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Людмила Александровна Шмелева - кандидат экономических наук, доцент, доцент кафедры операционного и отраслевого менеджмента факультета «Высшая школа управления»</p><p>Москва</p></bio><bio xml:lang="en"><p>Lyudmila A. Shmeleva - Cand. Sci. (Econ.), Assoc. Prof., Assoc. Prof. of the Department of Operational and Industry Management, Faculty of Higher School of Management</p><p>Moscow</p></bio><email xlink:type="simple">LyAShmeleva@fa.ru</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>Финансовый университет при Правительстве Российской Федерации</institution><country>Россия</country></aff><aff xml:lang="en"><institution>Financial University under the Government of the Russian Federation</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2025</year></pub-date><pub-date pub-type="epub"><day>20</day><month>11</month><year>2025</year></pub-date><volume>1</volume><issue>2</issue><fpage>6</fpage><lpage>17</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Зайцев Д.А., Прудников А.В., Хрипунова М.Б., Шмелева Л.А., 2025</copyright-statement><copyright-year>2025</copyright-year><copyright-holder xml:lang="ru">Зайцев Д.А., Прудников А.В., Хрипунова М.Б., Шмелева Л.А.</copyright-holder><copyright-holder xml:lang="en">Zaitsev D.A., Prudnikov A.V., Khripunova M.B., Shmeleva L.A.</copyright-holder><license xml:lang="ru" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>Данная работа распространяется под лицензией Creative Commons Attribution 4.0.</license-p></license><license xml:lang="en" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://www.digitarin.ru/jour/article/view/8">https://www.digitarin.ru/jour/article/view/8</self-uri><abstract><p>Мультимодальные чат-боты на платформе Telegram, управляемые оркестратором на базе большой языковой модели (LLM — Large Language Model), объединяют обработку текста, изображений и речи, расширяя привычные сценарии цифрового общения и закрывая дефицит естественного многоканального взаимодействия.Цель исследования — разработать и проанализировать архитектуру такого чат-бота, выявить ресурсные ограничения («экономику лимитов») и оценить их влияние на пользовательский опыт.Методы исследования. Проведен аналитический обзор решений 2023–2025 гг.; создан прототип бота (Python + Telegram Bot API) с LLM-оркестратором GPT‑4-класса, модулями компьютерного зрения, ASR/TTS и Retrieval-Augmented Generation. Экспериментальная выборка — 1500 запросов трех типов (текст, изображение, голос). Замерялись латентность, расход токенов, точность ответов и субъективная оценка пользователей (шкала SUS).Результаты. Оркестратор снизил средние затраты токенов на 41% за счет динамического выбора моделей и сжатия контекста; мультимодальные ответы подняли SUS-балл с 72 до 84; задержка ответа удержана в 6,8 с при 95-м процентиле. Гибридное хранилище знаний сократило число галлюцинаций на 36%.Выводы. Правильная LLM-оркестрация и продуманный учет лимитов (контекст, тарифы, скорость) позволяют существенно улучшить качество и надежность мультимодального Telegram-бота при контролируемых расходах; рекомендации применимы к корпоративным и публичным AI-ассистентам.</p></abstract><trans-abstract xml:lang="en"><p>Multimodal chatbots on the Telegram platform, orchestrated by a Large Language Model (LLM), fuse text, image and speech processing, filling the gap for natural multi-channel interaction.Purpose. To design and analyse such a chatbot architecture, identify resource constraints — “the economy of limits” — and evaluate their impact on user experience.Method. After a literature review (2023–2025) was created a prototype (Python + Telegram Bot API) based on GPT‑4-class LLM orchestrator, computervision, ASR/TTS and Retrieval-Augmented Generation modules. A test set of 1,500 queries (text, image, voice) was evaluated for latency, token cost, answer accuracy and user satisfaction (SUS scale).Results. Dynamic model routing and context compression cut average token expenditure by 41%; multimodal responses raised SUS from 72 to 84; 95th-percentile response time held at 6.8 s. A hybrid knowledge store reduced hallucinations by 36%.Conclusion. Well-designed LLM orchestration and efficient resource management (context window, pricing tiers, throughput) significantly enhance the quality and reliability of a multimodal Telegram bot while keeping costs under control; recommendations are transferable to both corporate and public AI assistants.</p></trans-abstract><kwd-group xml:lang="ru"><kwd>мультимодальный чат-бот</kwd><kwd>большие языковые модели</kwd><kwd>оркестрация</kwd><kwd>Telegram</kwd><kwd>ограничения контекста</kwd><kwd>пользовательский опыт</kwd><kwd>архитектура</kwd></kwd-group><kwd-group xml:lang="en"><kwd>multimodal chatbot</kwd><kwd>Large Language Model</kwd><kwd>orchestration</kwd><kwd>Telegram</kwd><kwd>context limitation</kwd><kwd>user experience</kwd><kwd>architecture</kwd></kwd-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Shen Y., Song K., Tan X., Li D., Lu W., Zhuang Y. HuggingGPT: Solving AI tasks with ChatGPT and its friends in Hugging Face. arXiv. 2023;2303.17580. DOI: 10.48550/arXiv.2303.17580</mixed-citation><mixed-citation xml:lang="en">Shen Y., Song K., Tan X., Li D., Lu W., Zhuang Y. HuggingGPT: Solving AI tasks with ChatGPT and its friends in Hugging Face. arXiv. 2023;2303.17580. DOI: 10.48550/arXiv.2303.17580</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Wu C., Yin S., Qi W., Wang X., Tang Z., Duan N. Visual ChatGPT: talking, drawing and editing with visual foundation models. arXiv. 2023;2303.04671. DOI: 10.48550/arXiv.2303.04671</mixed-citation><mixed-citation xml:lang="en">Wu C., Yin S., Qi W., Wang X., Tang Z., Duan N. Visual ChatGPT: talking, drawing and editing with visual foundation models. arXiv. 2023;2303.04671. DOI: 10.48550/arXiv.2303.04671</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Xu Y., Gao W., Wang Y., Shan X., Lin Y-S. Enhancing user experience and trust in advanced LLM-based conversational agents. Computing and Artificial Intelligence. 2024;2(2). DOI: 10.59400/cai.v2i2.1467</mixed-citation><mixed-citation xml:lang="en">Xu Y., Gao W., Wang Y., Shan X., Lin Y-S. Enhancing user experience and trust in advanced LLM-based conversational agents. Computing and Artificial Intelligence. 2024;2(2). DOI: 10.59400/cai.v2i2.1467</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">De Wynter A., Wang X., Sokolov A., Gu Q., Chen S-Q. An evaluation of large language model outputs: discourse and memorization. Natural Language Processing. 2023;4:100024. DOI: 10.1016/j.nlp.2023.100024</mixed-citation><mixed-citation xml:lang="en">De Wynter A., Wang X., Sokolov A., Gu Q., Chen S-Q. An evaluation of large language model outputs: discourse and memorization. Natural Language Processing. 2023;4:100024. DOI: 10.1016/j.nlp.2023.100024</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Li X., Zhang R., Xu X. Toolformer 2.0: self-augmenting large language models with multimodal tools. arXiv. 2024;2405.11223. DOI: 10.48550/arXiv.2405.11223</mixed-citation><mixed-citation xml:lang="en">Li X., Zhang R., Xu X. Toolformer 2.0: self-augmenting large language models with multimodal tools. arXiv. 2024;2405.11223. DOI: 10.48550/arXiv.2405.11223</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Kibkalo M., Shevtsov M., Gusev I. Gemini Ultra: multimodal generative model performance evaluation. In: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025). 2025. DOI: 10.18653/v1/2025.acl-main.219</mixed-citation><mixed-citation xml:lang="en">Kibkalo M., Shevtsov M., Gusev I. Gemini Ultra: multimodal generative model performance evaluation. In: Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025). 2025. DOI: 10.18653/v1/2025.acl-main.219</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Zhang Z. et. al. “It’s a Fair Game”, or Is It? Examining How Users Navigate Disclosure Risks and Benefits When Using LLM-Based Conversational Agents. 2024. URL: https://adalerner.com/ZhangCHI2024-FairGame.pdf DOI: 10.1145/3613904.3642385</mixed-citation><mixed-citation xml:lang="en">Zhang Z. et. al. “It’s a Fair Game”, or Is It? Examining How Users Navigate Disclosure Risks and Benefits When Using LLM-Based Conversational Agents. 2024. URL: https://adalerner.com/ZhangCHI2024-FairGame.pdf DOI: 10.1145/3613904.3642385</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
