Tobesoft Develops In-House AI Agent-Based Technical Support Service to Maximize Support Speed and Quality
AX-focused enterprise development platform company Tobesoft (CEO Kim Moranhee) announced on the 9th that it has independently developed an AI Agent-based internal technical support history search service and is actively advancing AI assetization and technology validation using accumulated in-house technical support data.
The project goes beyond simply improving search functionality by reconstructing long-term accumulated customer support and technical data into an AI-powered intelligent search system designed for practical use in real business environments. Through this initiative, Tobesoft aims to maximize the utilization of internal technical knowledge while gradually establishing a foundation for applying AI technologies to actual operations.
To overcome the limitations of conventional keyword-based matching systems, Tobesoft implemented RAG (Retrieval-Augmented Generation) technology by combining Large Language Models (LLMs) with embedding technologies. This enabled the company to build a semantic search system capable of understanding context and user intent.
When users describe technical issues in natural language, the AI analyzes the intent behind the query and automatically retrieves the most contextually relevant historical support cases. Search results are also designed to include reference document information and metadata, improving the accuracy of technical verification and follow-up responses.
In particular, the service has been developed as a pilot project with customized structures reflecting the unique operational characteristics and data types of each organization, including product services, customer support, and overseas support teams.
For the product service organization, which frequently operates in field environments, Tobesoft introduced a mobile-based AI search environment. By utilizing workflow automation tools, technical support histories from the sales management system are automatically extracted and vectorized, then combined with high-performance AI models to enable rapid technical responses using mobile devices even on-site. This environment allows staff to instantly review past support histories and site conditions in real time.
The customer support organization currently manages more than 65,000 support records accumulated over the past five years after converting them into vectorized data. As a result, similar inquiry cases that were previously difficult to locate through conventional search engine-based keyword searches can now be identified quickly and accurately through semantic-based search capabilities.
In addition, Tobesoft introduced a process that automatically extracts and analyzes conversations generated during customer calls and remote support sessions. The system evaluates whether appropriate responses were provided and registers related information into the platform, contributing to improved technical support quality and operational efficiency.
The overseas support organization also utilizes AI Agents to automatically analyze lengthy consultation records and structure them into easily understandable categories such as ▲inquiry overview ▲causes and countermeasures ▲solutions. The service additionally includes automatic language translation features based on user language environments, allowing overseas users in regions such as Korea and Japan to utilize the system without separate translation processes.
The development of this service is particularly significant in that it establishes the foundation for “AI data internalization” by transforming accumulated technical support and customer service data into forms suitable for AI learning and utilization. Through this initiative, Tobesoft plans to reduce response times for repetitive technical inquiries while gradually expanding company-wide AI utilization systems in the future.
Tobesoft stated that the AI Agent-based search service represents the first step toward transforming internally accumulated technical support data from simple records into practical AI assets that can be actively utilized in real business operations. The company added that it will continue advancing AI testing and various technology research initiatives to respond to the rapidly evolving AI environment while strengthening tangible technological competitiveness that customers can directly experience.