English 
Headwaters Inc., a company engaged in the AI solutions business, is pleased to announce the launch of its “Agentic Retrieval-Augmented Generation (AAG)” service, which maximizes the performance of large-scale language models (LLMs) through access to external data and tools.
Headwaters has expanded its lineup of LLM services for companies using the “Azure OpenAI Service,” building RAG (Retrieval Augmented Generation) systems tailored to the specific needs of each company, as well as supporting the development of a variety of AI agents, including contact center AI agents, station staff AI agents, migration AI agents, and in-vehicle edge AI agents, as well as proofreading AI and translation AI.
源流 has received many inquiries from companies interested in the early adoption of “AI agents,” and there is a strong demand for task-specific AI agents that can replace some tasks in combination with business data accumulated within the company.
こちらもお読みください： 日本触媒とNTTコム、熟練オペレーターから学習するAIで化学品製造工程を自動化
To meet these needs, Headwaters will provide the “Agentic RAG” service, which has already been adopted by major clients. “Agentic RAG” is a technology that combines multi-stage search and autonomous inference processes, a hybrid model of AI agents and RAGs.
これにより、Wordで作成された現場報告書や商品設計書、Excelで作成された発注書や商品リスト、データベースに蓄積された管理データなど、業務ごとに最適化されたAIエージェントの導入が可能になります。
また、AIエージェントに依頼するタスクを分割することで、特定領域知識の専門性や作業効率を向上させることができ、参照するデータソースを絞り込むことで、RAGの精度向上にも貢献します。
By providing the “Agentic RAG Orchestrator,” which autonomously references dedicated AI agents specialized for specific business tasks, “Agentic Design Pattern” planning, which serves as an overall design guideline for AI agents, support for conceptualizing the optimal Agentic RAG architecture to meet necessary requirements, and the “SyncLect AI Agent”, a multi-agent platform for quickly building secure Agentic RAGs, we will build an “Agentic RAG” environment with high autonomy and scalability in the customer’s Microsoft Azure environment.
ソース PRタイムズ
日本語