Domestic AI Agent Trends (August 13, 2026 Issue)
Update Date: 2026/8/13
Executive Summary
As of August 12, 2026, domestic AI agents have evolved from generating responses to executing business tasks. TerraSky's "mitoco Buddy" now operates Salesforce, SAP, and other systems through A2A and MCP integration, while also strengthening permission management. Hubble has developed iterative search for contract data and an AI Workbench, advancing autonomy in searching, analyzing, and document creation.
*The created article content was input into Gamma to automatically generate slides. If you find slides easier to view, please take a look here.
1️⃣ TerraSky "mitoco Buddy" directly operates business systems via A2A and MCP integration
📎 Source: TerraSky Co., Ltd. Press Release
TerraSky has updated its MCP-compatible AI platform, "mitoco Buddy," adding the A2A protocol and new MCP integration. From a chat, users can operate "mitoco Buddy Lingo" for meeting record scheduling, transcription, and summarization, and reflect the results in systems like Salesforce. For SAP, it allows for querying and executing business data, and for Azure Files, it enables folder creation and file saving. Furthermore, it includes "launcher permissions" that allow only agent activation and conversation, access control via Gateway, and prevention of overwriting files with the same name, simultaneously advancing functional expansion for operating external systems and a least-privilege design.
2️⃣ Hubble "Contract Flow Agent" provides an iterative search foundation and builds an "AI Workbench"
📎 Source: Hubble Inc. Press Release
Hubble has expanded the foundation of its contract AI agent, "Contract Flow Agent," and begun providing autonomous search and investigation functions targeting contract data. The AI interprets the user's objective, target, and conditions, and searches across contract text, comments, change history, ledger information, and similar documents. A key feature is that it does not stop after one attempt, but executes multiple times while changing search conditions based on the results. Additionally, it has built an "AI Workbench" where the AI executes programs in an isolated cloud environment for each request, handling document creation/revision, calculations, large-scale document analysis, and deliverable verification. Note that some uses, such as autonomous reflection and saving to files, are future development concepts.
Comprehensive Analysis
The features observed from the topics on August 12, 2026, indicate that the competitive axis for enterprise AI is shifting not only from "generation accuracy" but also to "connectivity with existing systems" and "mechanisms for safely entrusting business tasks." Moving forward, in addition to the standardization of A2A and MCP, practical operational foundations equipped with permission management and isolated environments are likely to become important.
Future Points of Interest
Attention will be on whether A2A and MCP become the standard for enterprise system integration. If standardization progresses, business automation across multiple AIs and SaaS platforms may accelerate.
How safely business tasks can be entrusted to AI will become critical. Governance functions such as least privilege, isolated execution, and operation logs are likely to influence adoption decisions.
