Muhammad Azkal Azkia, . (2026) PENGEMBANGAN SISTEM KONSULTASI BERBASIS WEB MENGGUNAKAN RETRIEVAL-AUGMENTED GENERATION (RAG) BERBASIS LANGGRAPH DI OMBUDSMAN RI. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.
![[img]](http://repository.upnvj.ac.id/style/images/fileicons/text.png) |
Text
ABSTRAK.pdf Download (226kB) |
|
Text
AWAL.pdf Download (1MB) |
|
Text
BAB 1.pdf Restricted to Repository UPNVJ Only Download (295kB) |
|
Text
BAB 2.pdf Restricted to Repository UPNVJ Only Download (607kB) |
|
Text
BAB 3.pdf Restricted to Repository UPNVJ Only Download (762kB) |
|
Text
BAB 4.pdf Restricted to Repository UPNVJ Only Download (3MB) |
|
Text
BAB 5.pdf Download (268kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (248kB) |
|
Text
RIWAYAT HIDUP.pdf Restricted to Repository UPNVJ Only Download (124kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository UPNVJ Only Download (5MB) |
|
Text
HASIL PLAGIARISME.pdf Restricted to Repository staff only Download (55MB) |
|
Text
ARTIKEL KI.pdf Restricted to Repository staff only Download (837kB) |
Abstract
The Ombudsman of the Republic of Indonesia (Ombudsman RI) is a state institution authorized to supervise public services under Law Number 37 of 2008. The high volume of complaints, totaling 10,846 Public Reports in 2024, along with reliance on conventional consultation methods creates obstacles including cost, limited technological access, and repetitive procedural explanations that reduce staff productivity. This research developed a web-based consultation system with an AI chatbot orchestrated by LangGraph StateGraph as the main framework managing the Retrieval-Augmented Generation (RAG) flow through five interconnected nodes supporting concurrent search execution, conversational state management, and dynamic routing based on intent classification. The system was developed using the Agile methodology with the Extreme Programming (XP) framework and equipped with Zoom consultation escalation, a Public Satisfaction Survey, and an admin dashboard. The pipeline utilizes Groq Llama 3.3 70B as the Large Language Model and ChromaDB containing 12 seed documents. Comparison on 15 test cases across 11 categories showed that the LangGraph+RAG model outperformed with an 80% pass rate and 82% in-context accuracy compared to Gemini without RAG (60% and 45%). Black Box Testing on seven modules reached 100%, while User Acceptance Testing involving 41 respondents obtained a weighted average of 4.558 out of 5.00 or 91.16% in the Very Good category.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Additional Information: | [No.Panggil: 2210511038] [Pembimbing 1: Indra Permana Solihin] [Pembimbing 2: Kharisma Wiati Gusti] [Penguji 1: Ridwan Raafi'udin] [Penguji 2: Ichsan Mardani] |
| Uncontrolled Keywords: | LangGraph, AI chatbot, RAG, Ombudsman RI, public service |
| Subjects: | Q Science > QA Mathematics > QA76 Computer software T Technology > T Technology (General) |
| Divisions: | Fakultas Ilmu Komputer > Program Studi Informatika (S1) |
| Depositing User: | MUHAMMAD AZKAL AZKIA |
| Date Deposited: | 21 Jul 2026 06:14 |
| Last Modified: | 21 Jul 2026 06:22 |
| URI: | http://repository.upnvj.ac.id/id/eprint/50963 |
Actions (login required)
 |
View Item |