Types of AI Infographic

Artificial intelligence (AI) is everywhere, permeating our work, lives, and conversations. Many of those conversations portray AI as one thing, but it’s actually a range of technologies. Understanding the differences is important to making good decisions about AI. As Colleen Jones notes in our recent report, What Makes Content Operations Successful in the Age of AI?,

You have AI options. Use them wisely. AI isn’t one-size-fits-all. What works for another organization might stall yours.

To add a bit of fun, this infographic compares different types of AI to droids and Jedis in Star Wars. The comparisons aren’t perfect, but they’re useful to understanding how machine learning is different from agents, for example. A text explanation of the infographic is available below the image.

Machine Learning = Droids Like R2D2

Why

• Highly trained to do very specific tasks
• Excellent within their narrow role
• Don’t “reason” broadly or improvise outside training
• The original that made more sophisticated droids possible

Hot Take

Don’t you love it when tasks are completed as expected? R2D2 will always be part of the team.

Examples

Recommendation engine (recommending similar products), sentiment analysis (classifying tone of feedback)

LLMs = Protocol Droids Like C-3PO

Why

• Understands and generates language
• Knows a lot, but only from training
• Sounds intelligent, but lacks timely real-world awareness

Hot Take

Interesting but risky in an enterprise thanks to that stochastic tendency. Endearingly polite. Mostly loyal.

Examples

Taking notes, writing copy, summarizing documents

RAG = C-3PO with Access to Jedi News + Archives

Why

• Still an LLM
• But now it can look things up in real time
• Reduces hallucination, increases accuracy

Hot Take

More interesting. Helps address risks of LLMs alone. That stochastic output, tho….

Example

Chatbot that answers using your internal docs or live data

Agent = Jedi on a Mission with Tools

Why

• Has goals
• Can plan steps
• Can use tools (APIs, databases, browsers)
• Can decide what to do next
• Remembers past actions and outcomes

Hot Take

Memory can address challenges with using AI across CX channels. Getting workflow right? A new challenge.

Example

AI that queries databases, sends emails, or posts content on a website

Agentic AI = A Jedi Council + Autonomous Field Jedi

Why

• Multiple agents collaborating
• Self-directed goal pursuit
• Long-term memory, planning and adaptation
• Minimal human intervention

Hot Take

Powerful. Let’s make sure one agent can complete a mission before giving it buddies. Please and thank you.

Example

AI that handles response to and escalation of customer service requests

May the AI and content force be with you.