Saltware Launches AI Data Platform 'Sapie-ADP'
IT DAILY ·
✦ AI Summary
Saltware announced on the 30th that it has officially launched 'Sapie-ADP (AI Data Platform),' an enterprise AI data platform.
'Sapie-ADP' is an integrated platform that links data connectivity and management, natural-language analysis, and agent-based work use.
The platform provides basis-checking functions, access-permission and security-policy-based controls, and support for network-segregated environments.
Saltware announced on the 30th that it has officially launched 'Sapie-ADP (AI Data Platform),' an enterprise AI data platform. 'Sapie-ADP' is a platform that consolidates the core technologies and functions of Saltware's existing AI, data, and security solutions.
'Sapie-ADP' is an integrated platform that supports enterprise data management and AI use. Users can query and analyze information in natural language based on distributed data inside the company, and they can build and use AI agents tailored to their own work.
The scope of enterprise AI use has recently expanded from document drafting and question answering to internal data analysis and task execution. As a result, the importance of distributed data connectivity and AI infrastructure based on business context is growing, and 'Sapie-ADP' was designed to respond to these changes.
The platform addresses related needs by linking data connectivity and management, natural-language analysis, and agent-based work use. It also provides a function for checking the basis for answers, as well as a function to control AI use based on a company's access permissions and security policies.
Sapie-Lake, which handles data connectivity and management, supports connections to heterogeneous databases. It also supports the use of unstructured data such as documents and drawings, and can be linked with existing ERP, DB, and object storage systems. This allows companies to build an AI environment without a complete data migration.
Sapie-Lake is based on ontology and a knowledge graph. In this framework, it defines and manages relationships between business terms and data. Based on this, it provides the foundation for AI to interpret and use data in line with a company's business context.
Built on this foundation is Sapie-DeepQ, a natural-language data query solution. Sapie-DeepQ supports easy querying and analysis of connected data. It converts user questions into SQL and performs the necessary data queries and analysis.
As a result, users can check the information they need for their work without having to write complex query statements themselves.
'Sapie-Agent' is an enterprise AI agent platform. It was designed to process user requests by using connected data and analytics functions. It provides work-related answers by linking data queries and analysis with retrieval-augmented generation (RAG)-based document question answering.
To this end, Sapie-Agent applies a multi-agent orchestration architecture. Multiple specialized agents share responsibilities and collaborate. They divide up tasks such as data queries, document searches, and analysis, and support the integration of each agent's results.
Customers can use the agents provided. They can also directly configure agents tailored to their own work. In this process, the system reflects company-specific data and work rules to create agents suited to each department and business purpose, and connects the tasks they need.
The platform can also be used in manufacturing. A single request can handle a query on production performance and a check of related process documents. Each agent searches for the necessary information in production data and process documents, and organizes the analysis results.
This makes it applicable to work that requires handling multiple sources of information together. In particular, it helps reduce the burden on manufacturing sites, where staff often have to collect materials across multiple systems.
Saltware has established a design that supports users in reviewing AI analysis results. The design is structured so that when reviewing AI results, users can check the basis and calculation criteria together.
The items provided are the basis data, reference date, applied formula, original source, and data update time. When the basis data is insufficient, the system marks it separately and also indicates areas that require additional confirmation.
At the same time, it links existing system user permissions and organization-specific access policies. This is intended to support the use of AI agents within the range of data users are allowed to access.
On the security side, it has applied the guardrail function of the AI security product Sapie-Guardian. It is designed to support control over AI use based on company security policies and AI usage policies.
It has also been configured to adapt to customer operating environments, and can be used in conjunction with various commercial AI models and local LLMs. It also supports network-segregated environments with limited internet connectivity, and in such environments it supports queries and analysis of internal company data based on local LLMs.
Saltware reflected the implementation experience it has accumulated in the manufacturing and defense sectors in this product. The experience reflected includes cross-analysis of heterogeneous databases, natural-language analysis linked to ERP, similarity search based on drawings, and question answering based on process knowledge.
Based on this, Saltware plans to expand supply, focusing on the manufacturing, defense, and public sectors. The target for expanded supply is the public and enterprise markets. The manufacturing, defense, and public sectors have the characteristic that data security and internal control are important.
Lee Jeong-geun, CEO of Saltware, said that if enterprise AI is to be used in real business operations, an agent environment capable of connecting enterprise data and carrying out data-driven work is necessary. He also explained that Sapie-ADP supports data connectivity, AI-based analysis, and business use in a single flow while maintaining existing data environments.
Saltware held a hands-on seminar on the 16th. The seminar included practical sessions on building an enterprise AI Gateway using LiteLLM and Amazon Bedrock, as well as operating Claude Code in an internally controlled environment.
Source: IT DAILY · Kwon Young-seok
Original: https://www.itdaily.kr/news/articleView.html?idxno=241918
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
View originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.