Skip to main content

Tasks

A Task is a specific unit of work that an agent needs to complete within a Job. It provides the concrete instructions, context, and expected output for an individual agent's contribution to the larger project.

What a task defines​

Tasks are the building blocks of any multi-agent workflow. Each task in Ordify specifies:

  1. Name & Description: A clear title and a detailed explanation of what the agent needs to do.
  2. Assigned Agent: The specific agent (e.g., "Researcher") that will perform the task.
  3. Expected Output: A description of what the final result should look like (e.g., "A 500-word summary in Markdown format").
  4. Tools: The specific tools required for this task.

Key features​

Inputs and outputs​

Tasks are designed to handle data. They take inputs (like a search query or a draft document) and produce a structured output that can be passed to the next agent in the job.

Context and dependencies​

In a sequential job, tasks can depend on each other. A "Drafting" task might depend on the context provided by a previous "Research" task — the output flows automatically.

Tasks vs. Skills​

It's important to distinguish between Tasks and Skills:

  • Skills are the training or capabilities an agent has (e.g., "Knows how to use Google Sheets").
  • Tasks are the assignments an agent is given (e.g., "Update the 'Inventory' sheet with the latest prices").

An agent uses its Skills to complete its assigned Tasks.

Creating and managing tasks​

Define tasks directly within the Agent Builder or the Jobs interface:

  1. Select the agent: Assign the task to the agent with the right expertise.
  2. Write the instructions: Be as specific as possible about the required steps and goals.
  3. Define the format: Specify whether the output should be a text summary, a JSON object, a file, etc.
  4. Add to job: Incorporate the task into your job's workflow.

Summary​

Tasks are the individual gears that drive a job's engine. By defining clear, well-structured tasks, you ensure that each agent understands its role and contributes effectively to the overall workflow.