Skip to main content
An Agent is an autonomous AI that decides which tools to use based on user input. Unlike traditional workflows where you define every step, agents make intelligent decisions about what actions to take.

​
Types of Agents in Markifact

Markifact offers three distinct types of agents, each designed for different use cases:

​
1. AI Agent Node

The AI Agent Node is a workflow component that brings AI-powered decision-making into your workflows. You connect various tools to this node, and the AI agent decides which tools to use based on the context and user input. Key characteristics:
  • Lives within your workflows alongside other nodes
  • You connect AI tools (GA4, Google Ads, Sheets, etc.) directly to it
  • Makes intelligent decisions about which connected tools to use
  • Ideal for building conversational interfaces and adaptive workflows
Use cases:
  • Slack chatbots that pull data from multiple sources
  • Interactive marketing assistants
  • Dynamic reporting systems
Learn more in the AI Agent node documentation.

​
2. Task Agent

The Task Agent operates outside of workflows and is designed for one-time, ad-hoc tasks. Unlike recurring workflows that run on schedules or triggers, Task Agents handle immediate, standalone requests. Key characteristics:
  • Works independently, not part of a workflow
  • Handles one-time tasks and ad-hoc requests
  • No recurring execution or scheduling
  • Perfect for quick tasks that don’t need workflow automation
Use cases:
  • Quick data analysis requests
  • One-off report generation
  • Immediate data lookups or transformations

​
3. Markifact Copilot

The Markifact Copilot is your AI assistant that helps you build workflows. Instead of manually creating and connecting nodes, you can describe what you want to accomplish, and the Copilot will construct the workflow for you. Key characteristics:
  • Builds workflows based on natural language descriptions
  • Understands your requirements and suggests the right nodes
  • Accelerates workflow creation
  • Helps both beginners and experienced users work faster
Use cases:
  • “Create a workflow that pulls Google Ads data and sends it to Slack”
  • “Build a report that combines GA4 and Meta Ads data into Google Slides”
  • Quickly prototyping new automation ideas

​
The Key Difference

​
Traditional Workflow

You build a fixed sequence: Trigger → Get Data → Process → Send Every input is predefined. The workflow always follows the same path.

​
Agent Workflow

You provide tools and instructions. The AI decides: What data to get? How to process it? Where to send it? The agent adapts to different questions and contexts dynamically. Slack Marketing Bot Example

​
Why Use Agents?

Perfect for unpredictable interactions:
  • Slack chatbots that handle varied questions
  • Customer support that needs different data sources
  • Research tasks where the path depends on findings
Not ideal for:
  • Predictable, repetitive tasks
  • When exact sequence matters
  • Performance-critical workflows

​
How They Work

  1. You provide instructions - “You’re a marketing assistant”
  2. You connect tools - GA4, Google Ads, Email, etc.
  3. User asks questions - “How did our ads perform?”
  4. AI chooses tools - Pulls Google Ads data automatically
  5. AI responds - With insights and analysis
The agent maintains conversation memory and builds context over multiple interactions.

​
Common Patterns

Marketing Assistant
  • Tools: GA4, Google Ads, Meta Ads, reporting
  • Platform: Slack, Teams
  • Handles: Performance questions, reporting requests
Data Analyst
  • Tools: Multiple data sources, analysis tools
  • Platform: Business dashboards
  • Handles: Ad-hoc analysis, insights generation

​
Getting Started

Ready to build? Check the AI Agent node documentation for detailed setup and configuration.