Skip to main content

Create an agent

Creating a custom AI agent in Ordify requires no coding. In just a few steps you can build an agent tailored to your specific needs and start using it immediately.

Step-by-step guide​

Step 1: Start creating your agent​

Click "+" on the Agent Navigation to open the agent creation modal, where you'll configure your agent's profile, capabilities, and knowledge.

New Agent Navigation

Step 2: Use Agent Builder​

Click the Agent Builder button to quickly generate your agent's core details.

Agent Builder Interface

The Agent Builder asks two key questions:

  1. "What problem should this agent solve, and what should happen when it succeeds?"

    • Describe the problem and the desired outcome
    • This generates your agent's Goal

  2. "What role should this agent have, and what instructions or guidelines should it follow in its behavior?"

    • Describe the role and behavioral instructions
    • This generates your agent's Role, Instructions, and Backstory

Based on your answers, the Agent Builder automatically fills in:

  • Goal: What the agent should accomplish
  • Instructions: How the agent should behave
  • Backstory: The agent's context and background
  • Description: A summary of the agent's capabilities

Visual configuration​

Once generated, you can fine-tune every field in real time:

  • Role: The agent's function (e.g., "Marketing Director," "Support Agent")
  • Goal: The objective the agent pursues
  • Backstory: Step-by-step behavioral instructions
  • Description: A short summary shown in the Library

Step 3: Add application integrations and skills (optional)​

In the Integrations tab, connect your agent to external applications.

Agent Integration Interface

Connect Apps:

  • Click Connect App
  • Browse and select from over 3,000 available applications
  • Connect popular services like Gmail, Google Calendar, Notion, and more

MCP Servers:

  • Click Add MCP Server
  • Connect custom MCP (Model Context Protocol) servers for internal tools
  • See MCP Servers for setup details

Skills:

  • Click Add Skill
  • Search or browse for existing skills in your Library
  • Attach Skills that bundle focused instructions and tools for specific workflows

Step 4: Add knowledge documents (optional)​

In the Knowledge tab, add documents to enable RAG (Retrieval Augmented Generation) for your agent.

Agent Knowledge Interface

RAG allows your agent to:

  • Access up-to-date information from your documents
  • Reference company knowledge, policies, or procedures
  • Answer questions based on your proprietary content

Knowledge configuration:

  • Select documents to enable RAG for context-aware responses
  • Upload new documents using the + Upload button
  • Select from your existing document library
  • Assign entire synced folders for automatic updates

Use the Document Library to manage RAG data, upload documents, organize folders, and share them with your team. For a full guide see RAG setup.

Step 5: Connect Jobs (optional)​

In the Jobs tab, you can allow this agent to invoke specific Jobs from within a conversation:

  1. Open the Jobs tab for your agent
  2. Select the Jobs this agent is allowed to call
  3. Update the agent's Instructions to describe when and how to invoke each job (for example, "When a user asks for a full proposal, call the 'Draft Sales Proposal' job with the account details from the conversation.")

See What is a job and Agents calling Jobs for more detail.

Step 6: Create and use your agent​

Click Create Agent to finish. Your agent is now ready to use.

Using your agent​

Chat directly with your agent​

  1. Click on the agent in the left sidebar to open a direct chat
  2. @mention the agent by name in any conversation

Use in multi-agent workflows​

Your agent can be added to Jobs — multi-agent workflows where multiple specialized agents collaborate on complex tasks.

What's next​

  • Understand Skills to give your agent specialized capabilities
  • Set up Knowledge to ground your agent in your own documents
  • Build a Job to have multiple agents work together