What raia Command gives you
- A Data Room file explorer that turns any document into agent-ready knowledge.
- Agent configuration: instructions, skills, channels, and AI settings.
- Live connectors to Zendesk, Salesforce, Google Drive, Jira, GitHub, SharePoint.
- Simulations and project search for QA and content discovery.
The Command Interface
Command is the operational command center of your AI workforce. It is organized around two primary workspaces: the Data Room (where knowledge lives) and Agent Configuration (where behavior is defined).
The Data Room
Looks and functions like a traditional file explorer, but every file uploaded is converted to Markdown and embedded into a Vector Store for semantic search.
- Smart Upload: Drag and drop and let AI classify and title files automatically.
- File support: PDFs, Word, Excel, PowerPoint, images, and raw data (JSON, CSV).
- Status tracking: Every file shows its exact position in the pipeline — Queued → Converting → AI Ready → Training.
Agent Configuration
- Instructions & guardrails: Persona, tone, boundaries, escalation rules.
- Skills & channels: Enable Live Chat, SMS, Webhooks, Functions, and more.
- AI settings: Choose whether the agent should automatically convert and train on new documents as they arrive.
Key Features & Workflows
| Feature | Description | Best for |
|---|---|---|
| Data Connectors | Pull content directly from Zendesk, Salesforce, Google Drive, Jira, GitHub, and SharePoint. | Keeping agent knowledge automatically synced with existing systems. |
| Generate Instructions | Select documents and let raia auto-draft a complete system prompt from their contents. | Quickly bootstrapping a new agent's core instructions. |
| Simulations | Run scripted, automated conversation tests against your agent before going live. | QA and regression testing after updating instructions. |
| Project Search | Full-text semantic search across the converted Markdown of every file in the Data Room. | Finding exactly where a specific policy or fact lives in training data. |
How to Get Started
- 01
Create a Project
Click 'New Project' in the sidebar to create a dedicated workspace and Data Room.
- 02
Upload Knowledge
Drag and drop your training documents, or connect an external system like Zendesk.
- 03
Configure Your Agent
In the Agents tab, define instructions and select the project as the knowledge source.
- 04
Test in Copilot
Before going live, chat with your agent in Copilot to verify its responses.
Related Training
Path
Functional Expert
Plan your agent, write instructions, build the knowledge base.
Path
Technical Builder
Advanced prompt engineering, webhooks, and custom functions.
Tip
The fastest way to learn Command is to upload a folder of real documents and watch the pipeline tick through Converting → AI Ready → Training in real time.