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

# AI Agent

> Ask questions about your inventory and sales in natural language, save prompts, and revisit past conversations.

The **AI Agent** is a conversational assistant. Ask about your inventory, sales, and
operations in plain language and get answers grounded in your organization's data, streamed
back as the agent works.

## What you can do

* **Chat** — ask questions ("which products are at stock-out risk this week?") and get
  **streaming** answers. The agent can call tools and show their results inline.
* **Organize by topic** — conversations are categorized (e.g. global insight, orders,
  stocks, sales, financial planning).
* **Prompt library** — save and reuse your best prompts (title, content, category, tags),
  so you don't retype common questions.
* **Archive** — archive a conversation to keep your list tidy; the archive is read-only and
  you can unarchive anytime.

## Privacy & scope

Your conversations and your prompt library are **private to you** within your organization —
other users don't see them. Each request carries your user and organization context, so the
agent only reasons over data your organization can see.

## Working with the agent

<Steps>
  <Step title="Open the agent">
    Go to **Apps → AI Agent** and type a question in the composer.
  </Step>

  <Step title="Refine in the thread">
    Follow up in the same conversation — the agent keeps context across the thread.
  </Step>

  <Step title="Save useful prompts">
    Add prompts you reuse to the **Prompt library**.
  </Step>

  <Step title="Tidy up">
    Archive finished conversations; find them later under **Archive**.
  </Step>
</Steps>

<Note>
  The agent complements the rest of Solya: it helps you find and understand, while you act
  through [plans](/en/inventory-plans/overview), [alerts](/en/signals/alerts), and
  [workflows](/en/automation/workflows). For programmatic agents, see the
  [MCP gateway](/en/mcp/overview).
</Note>