What Are AI Agents?
AI Agents:
- Act autonomously
- Learn from experience
- Make decisions
Agent Capabilities
- Perceive
- Interpret
- Plan
- Act
- Learn
Types of Agents
Simple Reflex
- If X → Do Y
Model-Based
- Uses memory
Goal-Based
- Optimises outcomes
Agents vs Assistants vs Bots
| Feature | Agents | Assistants | Bots |
|---|---|---|---|
| Autonomy | High | Medium | Low |
| Learning | Yes | Limited | No |
| Complexity | High | Medium | Low |
AI Agent Components
- LLM → Brain
- Prompts → Instructions
- Tools → Actions
- Knowledge → Data
- Memory → History
Make AI Toolkit
Capabilities:
- Text analysis
- Summarisation
- Translation
- Sentiment analysis
- Data extraction
AI Content Extractor
Works with:
- PDFs
- Images
- Audio
Functions:
- Extract text
- Describe images
- Transcribe audio
AI Agents in Make – Setup Process
- Create Scenario
- Add Trigger (Webhook/manual)
- Add “Run an Agent” module
- Add Tools
- Add Knowledge files
- Connect AI provider
- Write instructions
- Define settings
Agent Settings
- Speed vs Cost
- System prompts
- Input/output format
- Feedback loops
Tools & Knowledge
Tools
- APIs (Google Maps, Slack)
Knowledge
- Files with structured data
MCP (Model Context Protocol)
What is MCP?
A standard to connect AI with tools
→ “USB-C for AI”
Components
- Host → AI app
- Client → Connector
- Server → Tools
Benefits
- Dynamic tool discovery
- Simplified integration