AI CROSS launches MCP server enabling AI agents to control RCS/SMS messaging directly
The AMW Read
The event is an incremental product update for a known company in the AI agents subsegment; adds MCP integration to an existing platform but does not shift competitive dynamics outside Japan.
AI CROSS launches MCP server enabling AI agents to control RCS/SMS messaging directly
AI CROSS (AI CROSS株式会社), a Tokyo-based messaging and AI solutions provider listed on the Tokyo Stock Exchange Growth Market, has launched an MCP (Model Context Protocol) server for its enterprise RCS/SMS platform Zettai Reach! RCS (絶対リーチ！RCS). The server, released June 24, 2026, allows AI agents including Claude Desktop and GitHub Copilot to send, manage, and analyze RCS and SMS campaigns through natural language commands only, replacing manual API or scripting workflows.
Why it matters: This launch exemplifies the growing acqui-licensing pattern — not through M&A but via protocol-driven integration — where AI agents gain direct operational control over legacy enterprise communication channels. By embedding RCS/SMS functionality into the MCP ecosystem, AI CROSS effectively turns its platform into a tool-using agent capability, reducing friction for enterprise marketing and customer engagement teams. The move signals that MCP is becoming a distribution-standard for business messaging, reinforcing the hyperscaler-like network effects of agent-friendly protocols.
From an industry perspective, AI CROSS is a relatively small player in global messaging infrastructure, but its early support for agent-native control positions it ahead of larger SMS aggregators that still rely on traditional APIs or dashboards. The company has verified compatibility with Anthropic and Microsoft agent tools, creating a wedge into enterprise workflows where RCS adoption (particularly in Japan) is accelerating. If the pattern spreads, MCP servers will become a new competitive battlefield for enterprise SaaS vendors seeking to remain relevant in an agent-driven operational model.