> ## Documentation Index
> Fetch the complete documentation index at: https://agents.concierca.ch/llms.txt
> Use this file to discover all available pages before exploring further.

# Automate with AI (MCP)

> Your entire workspace is operable by AI: connect Claude, ChatGPT, or any MCP client and give it 136 governed tools across agents, missions, data, knowledge, and market intelligence.

Concierca AI is not just an app you click through — **the whole workspace is a server that any AI can operate.** Connect Claude, ChatGPT, Claude Code, or your own agent once, and it can read your data, run your agents, drive your missions, and pull market intelligence — all under the same permissions, departments, and governance you already use in the app.

<Note>
  This is the same platform, two front doors: **you** use the app; **your AI** uses the MCP server. Everything an AI does is scoped to what its credential may see, logged, and reversible.
</Note>

## What your AI gets

One connection exposes **136 governed tools** across every part of the platform:

<Columns cols={2}>
  <Card title="Your data" icon="table" href="/api/data">
    Query every business table — orders, customers, products, procurement, LinkedIn, ads — with filters, counts, and projections. Write rows where granted.
  </Card>

  <Card title="Agents" icon="robot" href="/api/agents">
    List, inspect, create, update, schedule and run agents — every run human-gated, every decision logged.
  </Card>

  <Card title="Missions" icon="diagram-project" href="/api/missions">
    Drive multi-step orchestrated workflows: validate, run, diagnose, pause/resume, compare versions and runs, read economics.
  </Card>

  <Card title="Knowledge & memory" icon="brain" href="/api/knowledge">
    Search and write durable company knowledge, episodic memories, and reusable skills.
  </Card>

  <Card title="Market intelligence" icon="chart-line" href="/api/vertical">
    The Vertical layer — curated brand cards, comparisons, benchmarks and know-how for your market.
  </Card>

  <Card title="Governance" icon="shield-check" href="/api/approvals">
    Human-in-the-loop approvals, guardrails, and per-tool grants — the AI can propose, but people decide.
  </Card>
</Columns>

## Connect in one step

Point any MCP client at your workspace endpoint and sign in:

```text theme={null}
https://app.concierca.ch/api/mcp/v1
```

Full setup for Claude and ChatGPT is on [Connect via MCP](/api/mcp). Prefer a terminal? The [CLI](/api/cli) wraps the same surface.

## What that unlocks

Once connected, a colleague — or an AI — can just ask, in plain language, and the AI finds the right data and tools on its own:

* *"Check my purchasing"* → the AI queries the procurement queue and tells you what to order.
* *"Why did last night's mission fail?"* → it reads the run trace and points at the failing step.
* *"Compare us to Venezianico"* → it pulls both brand cards from market intelligence.
* *"Update the returns SOP"* → it finds the existing document and edits it in place.

No custom prompt engineering, no guessing — the workspace tells the AI what exists.

## Built for AI, safely

The [**For AI Agents**](/api/tools) section is the complete operating manual an AI reads to use Concierca correctly:

* The [**Operating Contract**](/api/ai/operating-contract) — how a session starts and the rule to discover before guessing.
* [**What are you trying to do?**](/api/ai/routing) — every intent mapped to the exact tool sequence.
* [**All Tools**](/api/tools) — the full 136-tool catalog.
* Recipes for [**missions**](/api/ai/mission-recipes) and the [**vertical playbook**](/api/ai/vertical-playbook).

Writes prefer a **dry-run preview**, consequential actions pass a **human gate**, and every action is scoped and audited — so opening your workspace to AI never means losing control of it.