Smartsheet launches MCP Server with more than 40 tools for AI agents

Smartsheet's new MCP Server arrives as organizations look to let AI agents act on real work rather than just answer questions. The server provides structured context—row relationships, column types, cell history, and comments—which produces token-efficient payload sizes rather than raw API payloads. According to the announcement, this engineering decision reduces cost and improves latency at scale when running hundreds of AI-assisted queries across a project portfolio.

The integration paths target two distinct audiences. Business users, project managers, and operations teams can use Amazon Quick, an agentic AI workspace with Smartsheet as a pre-built connector, to query and update sheets through natural language. Developers can use Kiro, an agentic IDE, to reference project specs and update task statuses without leaving their coding workflow. Both paths expose the same Smartsheet toolset, so the choice depends on where the work gets done rather than on feature differences.

For regulated or enterprise environments, the security model builds on the same infrastructure and permissions layer as the Smartsheet REST API. Agents can only access data the authenticated user has permission to see, and administrators can toggle AI features globally or configure write-tool access by user group. The server does not store prompts, which may ease compliance reviews for teams concerned about data leakage through AI assistants.