Bots, Agents, and other AI Characters and Tools

A fun overview to clarify the different roles of AI tools

Content Generation Declaration:

Creative Concept: Willem Maritz

Writing: ChatGPT

Image Generation: ChatGPT

Editing: Willem Maritz

Are you confused about all the AI “characters” out there? Struggling to understand the difference between an AI Bot and an Agent?

Picture your workplace on a busy Monday morning: projects flying in, emails piling up, and everyone playing their role to keep the operation running. That scene is the perfect mental model for today’s AI landscape. By visualising these AI tools as familiar colleagues, with their own quirks, strengths, and limits, you’ll grasp when to “hire” each one in your teaching or learning design workflow, and why a blended team often beats relying on a single superstar.

1. LLM (ChatGPT or Gemini): “Clever Claire - The On-Call Specialist”

Clever Claire - The On-Call Specialist

Think of a large language model or ChatGPT as the senior expert whose brain is crammed with encyclopedias. You ping her only when you need an answer or a piece of writing, and she responds instantly with broad, general knowledge. She doesn’t remember the last project unless you remind her, and she won’t act unless you ask. Her power is depth of expertise, not initiative.

2. Custom Bot (OpenAI Custom GPT): “Eager Evan - The Trained Intern”

Eager Evan - The Trained Intern

A Custom GPT is that intern that’s eager to get things done, but lacks the confidence to make any decisions. You’ve handed him a binder titled “How we do things.” You embed permanent instructions, reference files, and permitted actions, so he answers exactly the way your course or company expects. He’s terrific for repetitive, well-scoped tasks, like creating lesson plans, rubric checks, or FAQ replies, but he still waits for you to tap him on the shoulder.

3. Automation Bot (Zapier Zap or Make scenario): “Reliable Rachel - The Competent Assistant”

Reliable Rachel - The Competent Assistant

This character is your rule-based assistant who patrols the office 24/7. You trained her once: “When an Evernote note is tagged Draft, send it to ChatGPT and file the reply in Google Docs”, and she now executes that playbook automatically whenever the trigger fires. She excels at predictable, linear chores across apps, but she neither improvises nor plans new routes if something unexpected happens.

4. Agent (N8N or Langchain): “Ambitious Andrew - The Autonomous Associate”

Ambitious Andrew - The Autonomous Associate

An agent is the ambitious graduate you hired last month. Give him a goal (“prepare a week-long micro-course on prompt engineering”) and he breaks it into steps, decides which tools to call (search, RAG, code-execution), asks follow-up questions if he’s stuck, and reports back only when the job is done. He reasons, replans, and copes with surprises without constant supervision.

5. Agent Manager (Langchain or Pipedream): “Collaborative Colleen - The Team Leader”

Collaborative Colleen - The Team Leader

Finally, meet the team lead who juggles multiple specialists. You define a high-level objective (“launch the entire new corporate communications campaign”), and she assigns the research agent, the copy-writing agent, and the Zapier bot to their parts, coordinates hand-offs, checks quality, and escalates only big decisions. Her super-power is delegation and conflict-resolution among agents with different skills, ensuring the whole ensemble marches toward your strategic goal. 

Just like all these familiar human characters in a team, AI tools also have their own strengths and weaknesses. Understanding them, you can assemble a super team of AI tools to collaborate and support your human team.

Have fun!