Welcoming AI as a 'Team Member' - Launch of 'AxMates', an AI Platform Bundling Internal System Integration, Permission Design, and Hands-on Support
LifePrompt Inc. has launched 'AxMates', an AI platform that integrates with internal corporate tools via API to autonomously execute actual business tasks. It features strict data permission control using MCP and offers hands-on implementation support.
📋 Article Processing Timeline
- 📰 Published: March 29, 2026 at 21:00
- 🔍 Collected: April 1, 2026 at 13:27 (64h 27m after Published)
- 🤖 AI Analyzed: April 21, 2026 at 09:57 (476h 29m after Collected)
LifePrompt Inc. (Headquarters: Shinjuku-ku, Tokyo; CEO: Satoshi Endo) will launch 'AxMates', an AI platform where AI connects with internal systems to drive tasks to completion like a team member, starting April 1, 2026.
Unlike general-purpose chat AI such as ChatGPT, AxMates connects with internal tools (Google Drive, Slack, freee, HERP, etc.) via API to handle actual operations like information gathering, drafting, transcribing, and executing recurring tasks, rather than just conversing. For implementation, we have adopted a 'hands-on' provision format where we collaborate with customers on permission design—determining what data the AI can access and to what extent it can execute tasks.
---
Background: Why AxMates is Needed Now
With the spread of large language models (LLMs) like ChatGPT, many companies are working on utilizing AI in their business operations. However, the following challenges have been voiced from the front lines.
1. Simply asking questions in chat doesn't finish the work
While general-purpose AI is good at answering questions, it cannot access internal systems and operate within actual business workflows. Ultimately, manual work remains where humans must copy the AI's answers and transcribe them into other tools.
2. Workflow-type tools are prone to becoming technical debt
Workflow automation tools like n8n and Zapier are powerful instruments for their creators, but they easily become dependent on specific individuals, making maintenance difficult if key personnel leave. Furthermore, workflows force AI to follow fixed, human-defined flows. Thus, even as AI evolves, past definitions become bottlenecks, hindering continuous improvement.
3. Preparations to let AI touch internal data have not kept pace
While the introduction of AI agents is expanding, governance preparations such as 'what data to grant access to', 'how far to allow automatic execution', and 'how to protect confidential information' remain untouched in many companies.
AxMates was designed to comprehensively solve these challenges.
---
Features of AxMates
1. Connects with internal systems and drives work to completion
Integrates via API with internal tools such as Google Drive, Slack, Teams, freee, HERP, and Notion. For example, if you request in chat to 'summarize last month's sales,' it can collect related data from multiple tools and execute up to creating a draft in a specified format.
2. Holds business logic as natural language knowledge
Unlike workflow-type tools, it saves business procedures and rules as natural language knowledge (SOP). It can be checked and modified without programming knowledge, making handovers easy even when personnel change. As AI becomes smarter, it can perform higher-quality work from the same knowledge, preventing it from becoming technical debt.
3. Permission design using MCP (Model Context Protocol)
The system uses MCP to control what internal data the AI can access, what processing it can perform, and where it can output. It allows control at the processing and output levels, not just access rights, such as 'can output aggregated results but not individual data' or 'will not provide answers containing confidential information in this channel.'
4. Responds as the same 'single entity' from multiple channels
AxMates can be accessed through multiple routes, including Web chat, Slack, Microsoft Teams, email forwarding, Chrome extensions, and scheduled starts. No matter which channel you speak from...
Unlike general-purpose chat AI such as ChatGPT, AxMates connects with internal tools (Google Drive, Slack, freee, HERP, etc.) via API to handle actual operations like information gathering, drafting, transcribing, and executing recurring tasks, rather than just conversing. For implementation, we have adopted a 'hands-on' provision format where we collaborate with customers on permission design—determining what data the AI can access and to what extent it can execute tasks.
---
Background: Why AxMates is Needed Now
With the spread of large language models (LLMs) like ChatGPT, many companies are working on utilizing AI in their business operations. However, the following challenges have been voiced from the front lines.
1. Simply asking questions in chat doesn't finish the work
While general-purpose AI is good at answering questions, it cannot access internal systems and operate within actual business workflows. Ultimately, manual work remains where humans must copy the AI's answers and transcribe them into other tools.
2. Workflow-type tools are prone to becoming technical debt
Workflow automation tools like n8n and Zapier are powerful instruments for their creators, but they easily become dependent on specific individuals, making maintenance difficult if key personnel leave. Furthermore, workflows force AI to follow fixed, human-defined flows. Thus, even as AI evolves, past definitions become bottlenecks, hindering continuous improvement.
3. Preparations to let AI touch internal data have not kept pace
While the introduction of AI agents is expanding, governance preparations such as 'what data to grant access to', 'how far to allow automatic execution', and 'how to protect confidential information' remain untouched in many companies.
AxMates was designed to comprehensively solve these challenges.
---
Features of AxMates
1. Connects with internal systems and drives work to completion
Integrates via API with internal tools such as Google Drive, Slack, Teams, freee, HERP, and Notion. For example, if you request in chat to 'summarize last month's sales,' it can collect related data from multiple tools and execute up to creating a draft in a specified format.
2. Holds business logic as natural language knowledge
Unlike workflow-type tools, it saves business procedures and rules as natural language knowledge (SOP). It can be checked and modified without programming knowledge, making handovers easy even when personnel change. As AI becomes smarter, it can perform higher-quality work from the same knowledge, preventing it from becoming technical debt.
3. Permission design using MCP (Model Context Protocol)
The system uses MCP to control what internal data the AI can access, what processing it can perform, and where it can output. It allows control at the processing and output levels, not just access rights, such as 'can output aggregated results but not individual data' or 'will not provide answers containing confidential information in this channel.'
4. Responds as the same 'single entity' from multiple channels
AxMates can be accessed through multiple routes, including Web chat, Slack, Microsoft Teams, email forwarding, Chrome extensions, and scheduled starts. No matter which channel you speak from...