Learn
AI, GenAI and agents, explained with one task
Same job, three kinds of AI. Click through to see what each does and what’s left for you.
The explainer
Example task A customer emails asking whether you can make a gluten-free birthday cake for Saturday.
Plain AI Recognizes and sorts
Spots patterns in data. It labels, ranks or predicts, but it doesn’t write anything new.
- What it does
- Your email inbox flags the message as a customer inquiry and files it under “Orders”.
- Your part
- You still read it, check your calendar and write the reply.
- Familiar examples
- Spam filters, photo search, map routes
GenAI Creates content when asked
Generates new text, images or code from your prompt. It answers once, then waits for you.
- What it does
- You paste the email into a chat assistant and ask for a friendly reply. It writes a draft in seconds.
- Your part
- You copy the draft, check Saturday’s orders, fix details and send it.
- Familiar examples
- Chat assistants, image generators
Agentic AI Works through steps toward a goal
Takes a goal, plans steps, uses tools you’ve connected (like email or a calendar) and keeps going. Good ones check with you first.
- What it does
- Your helper sees the new email, checks Saturday’s calendar, reads your gluten-free notes and drafts a reply with a price from your price list.
- Your part
- You review the draft and press send. It never sends on its own.
- Familiar examples
- Assistants with tools, routines and memory
At a glance
| Plain AI | GenAI | Agentic AI | |
|---|---|---|---|
| Main job | Sort, label, predict | Write or make things | Carry out steps toward a goal |
| Starts when | Built into an app | You ask | You ask, or on a schedule you set |
| Uses your tools | No | Usually not | Yes, the ones you connect |
| Your role | Do the work | Edit and send | Set the job, approve the drafts |
Quick self-check
Five questions. Instant feedback.
Next step
See what an agent could do for someone like you.