01
Model
The engine that generates. ChatGPT, Claude, Gemini are products wrapped around models.
Lesson 01 / 07
Model · Tool · Agent · ~8 min
What Gen AI actually is
Separate models, tools, and agents — so you stop buying demos and start buying outcomes.
Not magic. A pattern completer.
Generative AI predicts the next useful chunk of text, image, or code from patterns it has seen. It does not “know” your company unless you give it the right context.
That is why the same model can draft a brilliant email one minute and invent a fake policy the next. Capability and reliability are different questions.
01
Model
Predicts the next useful chunk
02
Tool
A job wrapped around a model
03
Agent
Steps + tools + human gates
Fig. 01 — Three words teams confuse constantly
Keep these distinct
02
Tool
Software that uses a model for a job — summarise a call, score a lead, rewrite a brief.
03
Agent
A system that takes steps: read data, call tools, wait for a human, then act.
04
Workflow
The operating path across people and systems — where agents earn their keep.
Why this matters in a company
Buying “AI” without naming the shape leads to chat windows nobody uses, or automations that email customers without permission.
UP-ON’s work starts by diagnosing the workstream — then choosing chat, template, or agent on purpose.
Check
Your sales team wants “an AI that follows up with stalled deals in HubSpot and only pings a human when the deal is big.” What are they really asking for?
Check
Which statement is most accurate?