AI Integration (MCP) — Connect AI Assistants to Your Workspace
A built-in MCP server that gives AI assistants full context across chat, docs, tasks, and CRM. Not a chatbot. A protocol-level integration.
Most "AI integrations" in workspace tools are chatbots that can answer basic questions about your data. They run in a widget, they have limited access, and they cannot actually do anything. You still have to switch to the app, find the right page, and make the change yourself.
Zioan takes a different approach. It includes a built-in MCP (Model Context Protocol) server that gives AI assistants direct access to your entire workspace. Your AI assistant can search documents, create and update tasks, send chat messages, manage CRM contacts, schedule calendar events, and more. It does not just read your workspace. It works in it.
This works with Claude, ChatGPT, Gemini, and any AI tool that supports the MCP protocol. The AI runs on your machine and connects to your Zioan server, so workspace data stays on your infrastructure. The AI assistant does not store your data. It accesses it through the MCP connection when you ask it to.
What's included
- Built-in MCP server (no external plugins needed)
- Works with Claude, ChatGPT, Gemini, and any MCP-compatible AI
- Search across docs, snippets, tasks, and chat
- Create, update, and complete kanban tasks
- Send and read chat messages
- Create and edit documents and snippets
- Manage CRM contacts and follow-ups
- Schedule calendar events
- Check notifications and unread counts
- Built-in help system for onboarding guidance
- API token authentication with user-level permissions
- Three-step setup (generate token, install, configure)
How AI integration works in Zioan
What the AI can do
Through the MCP server, an AI assistant can perform real actions in your workspace:
- Search across documents, snippets, tasks, and chat messages
- Create, update, and complete tasks on kanban boards
- Send messages in channels and direct messages
- Create and edit documents and code snippets
- Manage CRM contacts, add notes, track follow-ups
- Schedule calendar events and set reminders
- Check notifications and unread messages
- Assign team members to tasks
Three-step setup
Generate an API token in your Zioan settings. Install the MCP server configuration in your AI tool. Point it at your Zioan instance. That is the entire setup. No OAuth flows, no webhook configurations, no third-party connector platforms.
Works with multiple AI tools
Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), and any MCP-compatible assistant can connect to the same MCP server. Your team members can use whichever AI tool they prefer. The integration is at the protocol level, not tied to a specific vendor.
Built-in help system
The MCP server includes a comprehensive help system that teaches the AI assistant every feature, screen, and workflow in Zioan. When a team member asks their AI assistant "how do I create a kanban board?" or "where are my follow-ups?", the AI can give step-by-step answers with full UI awareness. It knows where buttons are, what menu items exist, and how workflows connect.
Full workspace context
Because the MCP server connects to the same database that powers chat, docs, tasks, CRM, and calendar, the AI assistant has access to everything in context. Ask it to "find the document we discussed in the design channel last week and create a task for the follow-up items" and it can do exactly that, across multiple features in a single request.
API token authentication
Each user generates their own API token. The AI assistant operates with the same permissions as the user who generated the token. It cannot see spaces, documents, or contacts that the user does not have access to. Standard permission boundaries apply.
Why teams choose Zioan for AI integration
AI that actually does things
Most workspace AI features are limited to search or summarization. Zioan's MCP integration lets the AI assistant create tasks, send messages, update the CRM, and schedule events. It is the difference between "here is what I found" and "done, I created the task and assigned it to Alex."
Onboarding through conversation
New team members can ask the AI assistant how any feature works. The built-in help system means the AI knows every screen, every button, every workflow. Instead of reading documentation or asking colleagues, new hires can ask "how do I share a document with a client?" and get a step-by-step answer immediately.
Works across your entire workspace
A single prompt can span chat, docs, tasks, and CRM. "Summarize what happened with Client X this week: check their CRM notes, any tasks we completed, and messages in their channel." The AI pulls information from multiple features because they share one database. This is not possible with separate tools connected through a third-party integration layer.
No data sent to third parties
The MCP server runs on your Zioan instance. The AI assistant connects to it from your local machine. Workspace data is accessed live through the API, not copied to an external service. When the AI session ends, no workspace data remains outside your infrastructure. This is privacy-first AI integration.