Syahputra, Muhammad Harvian Dito (2026) Pengembangan Chatbot Berbasis Multi-Agent AI dan Agentic Retrieval-Augmented Generation untuk Pelayanan dan Sistem Rekomendasi di Puspaga Kota Probolinggo. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5027221039-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5027221039-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Digitalisasi pelayanan publik menuntut layanan yang responsif dan dapat diakses kapan saja. Puspaga Kota Probolinggo menghadapi berbagai kendala pada pelayanan administrasi, penjadwalan terapi anak, serta pemberian informasi yang masih dilakukan secara manual melalui WhatsApp dan layanan tatap muka. Kondisi tersebut menyebabkan keterbatasan layanan di luar jam operasional serta meningkatnya beban kerja administratif. Penelitian ini mengembangkan chatbot berbasis Multi-Agent AI dan Agentic Retrieval-Augmented Generation (Agentic RAG) untuk mendukung layanan informasi, penjadwalan terapi anak, konsultasi awal, dan sistem rekomendasi layanan pasien. Sistem terintegarasi dengan WhatsApp, back-end REST API, dan website dashboard, serta menerapkan arsitektur Multi-Agent AI yang terdiri atas Router Agent, Booking Agent, Anamnesa Agent, dan Summarization and Recommendation Agent yang terintegrasi dengan Agentic RAG. Digunakan konfigurasi nilai chunk size 1024, embedding model intfloat/multilingual-e5-base, dan LLM GPT-5.4. Evaluasi dilakukan menggunakan framework Ragas, pengujian Success Rate dan Response Time, serta User Acceptance Test (UAT). Hasil pengujian menunjukkan bahwa sistem memperoleh skor Ragas keseluruhan sebesar 0,779. Pengujian fungsional menghasilkan Success Rate 100% pada uji coba klasifikasi intensi dan fitur reservasi terapi anak, serta 83,33% pada fitur konsultasi awal. Rata-rata Response Time sistem adalah 12,95 detik. Hasil UAT menunjukkan tingkat penerimaan pengguna sebesar 4,6 dari 5 pada UAT Chatbot dan Website Dashboard.
==============================================================================================================================
The digitization of public services demands responsive services that are accessible at any time. The Probolinggo City Public Health Center faces various challenges in administrative services, scheduling children’s therapy, and providing information, which are still handled manually via WhatsApp and in-person services. These conditions result in limited service availability outside of operating hours and an increased administrative workload. This research develops a chatbot based on Multi-Agent AI and Agentic Retrieval-Augmented Generation (Agentic RAG) to support information services, child therapy scheduling, initial consultations, and a patient service recommendation system. The system is integrated with WhatsApp, a back-end REST API, and a website dashboard, and implements a Multi-Agent AI architecture consisting of a Router Agent, Booking Agent, Anamnesis Agent, and Summarization and Recommendation Agent integrated with Agentic RAG. The configuration uses a chunk size of 1024, the intfloat/multilingual-e5-base embedding model, and the GPT-5.4 LLM. Evaluation was conducted using the Ragas framework, testing Success Rate and Response Time, as well as a User Acceptance Test (UAT). Test results show that the system achieved an overall Ragas score of 0.779. Functional testing yielded a 100% Success Rate in the intent classification and pediatric therapy reservation feature trials, and 83.33% for the initial consultation feature. The system’s average Response Time was 12.95 seconds. UAT results showed a user acceptance rating of 4.6 out of 5 for the Chatbot and Website Dashboard UAT.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Chatbot, Multi-Agent AI, Agentic Retrieval-Augmented Generation, Large Language Model, Agentic AI |
| Subjects: | T Technology > T Technology (General) > T58.62 Decision support systems |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis |
| Depositing User: | Muhammad Harvian Dito Syahputra |
| Date Deposited: | 16 Jul 2026 08:04 |
| Last Modified: | 16 Jul 2026 08:04 |
| URI: | http://repository.its.ac.id/id/eprint/135226 |
Actions (login required)
 |
View Item |