L1-M1 / lesson 1 of 2

Understand what an AI system does

Learn what AI can do, what it needs, and what you still need to check.

10–15 min · 0/4 steps complete

Start here / A workplace decision

One request, three system configurations

Mira needs a summary of a supplier offer. Compare what the assistant can support with a prompt, an approved source and a calculator.

The assistant receives only a request to summarise the Northline offer.

Before you read / Think it through

In “One request, three system configurations”, what would you check or do first?

Pause and form your answer, then compare the reasoning. This reflection is not scored.

Compare the reasoning

Without the offer, it has no verified fee or payment terms. Fluent text does not supply missing evidence.

The ideas below give you a method for this decision. Then work through the rest of this case and try the decisions yourself.

Part 1 of 4 Learn the ideas

Choose a topic0 of 5 read

Recognise AI in everyday work

AI can help a system predict, rank, recommend or generate something. Start by asking what the tool actually does. A spreadsheet adding invoice totals follows a fixed calculation; a model estimating which invoices may be fraudulent uses patterns learned from data.

The distinction matters because an estimate needs a different kind of checking from a fixed calculation. A prediction is a judgement about an input, not a guaranteed fact. A product name or an “AI-powered” label is not enough to understand that judgement.

At your desk

Your inbox tool puts a customer message in the “urgent” queue. Ask what evidence it uses and what happens if it gets the priority wrong.

Remember this

Describe the task and its consequences before deciding how much to trust the tool.

The workplace case

MEET MIRA / YOUR CASE THROUGHOUT THIS MODULE

Mira is preparing a supplier summary for a client meeting. Her assistant produces a polished draft, but she has not verified the claims or whether this account is approved for client data. You are helping her decide what happens next.

Quick reference / five useful terms

Model: The learned component that helps produce predictions or generated content.
System: The working product: model, interface, connected sources, tools and controls.
Token: A piece of text, such as a word fragment or punctuation, used in language generation.
Grounding: Supplying relevant source material so a response can be based on it. Verification is still needed.
Hallucination: A plausible but unsupported or incorrect output. Also look for missing conditions.