Lesson 2

Making AI work for you

Where you are with AI today, and the problems worth handing to it

Personal utility

Where we are

Lesson 2 of 12: AI for you

01 · Done

Introduction to Generative AI

Mon 9 Nov · 08:40, 10:20

02 · You are here

AI for you: making it work in your day

Tue 10 Nov · 08:40, 10:20

03

Getting the most out of your AI tools

Wed 11 Nov · 08:40, 10:20

04

How AI systems work

Wed 18 Nov · 14:20

05

AI at work and in teams

Wed 18 Nov · 16:20

06

Spotting AI opportunities

Fri 20 Nov · 08:40, 10:20

07

Practical business with AI

Mon 23 Nov · 08:40, 10:20

08

Ethical AI and what AI costs

Fri 27 Nov · 14:20, 16:20

09

Building with AI

Mon 30 Nov · 12:40, 14:20

10

Build and pitch prep

Fri 4 Dec · 08:40, 10:20, 12:40, 14:20

11

Final presentations

Tue 8 Dec · 08:40, 10:20

12

Reflection and final Q&A

Wed 9 Dec · 08:40, 10:20

Recap

In lesson 1, our main objective was to:

Understand what AI is, why it matters now and how the tools around it are changing

What we covered

Why AI, why now

AI is spreading faster than any technology before it.

AI is not magic

It runs on data and rules people design, so it needs your judgment.

The main types of AI

Generative AI is one type among several: vision, language and prediction.

The AI landscape

Chatbots now reach into coding, copilots and agent builders.

Tools change fast

The principles last: the problem, asking and checking, judgment, people.

From Lesson 1

The AI landscape today

Where we stopped: the tools now, and what stays the same

AI assistants in context

There are more ways to interact with AI than ever before

Human-in-the-loop

Low level of human input

Standardised process

High complexity / variance

Automation

Automates a task in the background on set rules

Tools

Automate a task on demand by the team

Assistants

Support team tasks and opinions for thinking

AI agents

Autonomously complete a task

Copilots

“What would you like me to help you with?”

What stays the same

Tools change fast. The principles don’t.

Changes every
few months

Tools and appsModelsFeaturesPricesWho leads the market

Learn them as you go.

Changes slowly: what this course focuses on

The problem you’re solving

Start from the need, not the tool.

How you ask and check

Clear context in. Careful checking out.

Judgment and responsibility

You own the result: accuracy, privacy and fairness.

People and change

Habits and trust change slowly. Bring people with you.

Lesson 2 · Part 1

Your current state

How you use AI today, and what it is really like

AI is a superpower that can be used for almost anything

…right?

Reality check

False expectations vs reality

What we expect

The reality

It is always right

It can make things up and sound sure (hallucination)

Same question, same answer

Answers change each time you ask

It is neutral

It can repeat the bias in its training data

It thinks like a person

It predicts likely words. Logic can still be wrong

What I type stays private

Some tools learn from your chats unless you switch it off

It is free

Every answer uses energy and water

The reality

AI as the great leveller

It lifts those who start lower the most…

…but it fails quietly on tasks it cannot do.

758 consultants at Boston Consulting Group worked on real tasks, with and without GPT-4.

+43%

Lower half

Quality of work rose 43% for those who scored lowest on their own.

+17%

Top half

The strongest performers gained much less.

Source: Dell’Acqua et al., Navigating the Jagged Technological Frontier, Harvard Business School, 2023 (opens in a new tab)

Dotted circle: how hard tasks seem to us

What AI can really do

Brainstorm 20 names

Summarise a long report

Draft an essay plan

Count the letters in a word

Quote a real source exactly

Spot one small clue in data

The reality

The jagged frontier

AI is brilliant at some tasks and poor at others that look just as hard. You only find out which by testing.

−19

percentage points

On a task just outside the line, consultants with AI were less likely to get it right than those without.

So check its work, and learn where the edge is.

Source: Dell’Acqua et al., Harvard Business School, 2023 (opens in a new tab)

Lesson 2 · Part 2

Your desired state

What would you hand to AI if you could?

Score your sticky notes

Good problems for AI

Does it happen often?

Daily or weekly beats once a year.

Does it follow a pattern?

The same steps, or the same kind of answer, each time.

Is it mostly words or information?

Text, numbers, files and messages are where AI is strongest.

Does it take long by hand?

The more time it eats, the more you win back.

Can you check the result easily?

You stay responsible, so you need to spot mistakes.

More yeses, better problem. Pick your top three.

Keep these for yourself

Bad problems for AI

It must be 100% right

Medicine doses, legal forms, answers you hand in.

It uses private information

Other people’s details, passwords, anything confidential.

It needs information AI cannot see

Today’s news, what was said in class, what is only in your head.

It is about people and feelings

A hard message to a friend, a decision about someone.

It is quicker to just do it

A two-minute job takes longer to explain than to do.

Any of these? Move that sticky note to one side.

Before you hand it over

“Every time we hand work to a wizard, we lose a chance to develop our own expertise, to build the very judgment we need to evaluate the wizard’s work.”

Ethan Mollick, Wharton School

Activity · 5 minutes

What do you hate doing?

  1. Write down five things you hate doing in daily life.
  2. One per sticky note.
  3. Be specific: not “admin” but “finding the form for my visa”.
Replying to the group chat
Finding that one PDF
Planning what to cook
Remembering deadlines
Booking appointments
Writing the same email again

Which AI use case fits?

Match each problem to what AI can do

Your problemWhat AI can doWhere the information lives
Writing the same email againCreate…Your inbox
Catching up on a long readingExplain and summarise…Course files
Remembering deadlinesExtract…Email and calendar
Planning what to cook this weekReason and plan…Your head, and the fridge
Finding that one PDFExtract…Files and email

Your turn: fill in this table for your top three.

Identifying data sources

Where does the information live?

For each of your top three problems, where is the information needed to solve it?

AI can only help with what it can see.

Your head

Group chats

Email

Calendar

Notes

Files and Drive

AI memory

Browser tabs

Memory

But surely this is what memory solves?

It remembers you from chat to chat.

Memory

Example

Reference saved memories

Reference chat history

Saved memories

Studies entrepreneurship at ISM

Prefers short answers in plain English

English is their second language

Working on a business plan for a café

So what’s the risk?

Your memory lives in their tool.

Pay, or start again from zero.

“Old news. ChatGPT already remembers me.”

Years of ChatGPT

Every chat you have had
Your preferences and style
Your projects and context
Price × 10

All your context is inside. You are stuck, with no real choice as an end user.

Interoperability

Not a problem until it is

Before

A different plug for each

After

USB-C: one plug

The same is happening with AI memory.

Independent systems mean your choice and your control.

What if it is on your computer?

Your files are on your own machine.

“Just tidy my folders and that horrible desktop.”

Surely AI can fix this, right?

FileEditView

New folder (7)

invoice (1).pdf

Screenshot 2026-03-14 at 09.41.png

final_FINAL_v2.docx

Downloads (1)

budget copy.xlsx

Untitled(3).pdf

IMG_4021.HEIC

cv_old.docx

zoom_0.mp4

New folder

scan0012.pdf

essay draft

Screenshot 2026-04-02.png

Untitled

stuff

ticket.pdf

notes.txt

IMG_4022.HEIC

contract_signed(2).pdf

Document 1

old

Oh, what happens when it goes wrong?

01

When AI goes awry with its permissions

02

When you give more control than you believe you have

My Drive

Pitch deck.pptx

Contracts

Receipts 2026

Team photos

Trash

So what does that mean?

To help you, AI needs access

Your AI app

Group chats

Email

Calendar

Files and Drive

Your desktop

Notes

Browser tabs

Your head

No way in: you type it

…until Neuralink

Each line is your data, going to someone else’s computer.

What’s the risk with this?

Privacy

It can leak to others

If your chats train the AI, your details could turn up in someone else’s answers.

Other people’s details

Your inbox and chats hold your friends’ details too. They never agreed.

Security

More ways in

Every app you connect is another door if your account is hacked.

AI can be tricked

A hidden line in an email or web page can tell your AI to do the wrong thing.

Ethics

Who is responsible?

If AI sends the wrong message for you, it is still your message.

Is it fair?

When AI sorts or judges people, it can repeat the bias in its data.

We come back to each of these later in the course.

Recap

In this lesson, our main objective was to:

Make AI work for you, starting with your own problems

What we covered

What stays the same

Tools change fast. Start from the need, not the tool.

Expectations vs reality

AI is brilliant at some tasks and poor at others that look just as hard.

Good problems for AI

Pick tasks that repeat, follow a pattern and are easy to check.

Where the information lives

AI can only help with what it can see. Keep your memory where you control it.

The risks

Giving AI access brings risks to privacy, security and ethics.

End of lesson 2

Thank you

See you tomorrow

Next: Lesson 3 · Wednesday 11 November · 08:40

Getting the most out of your AI tools

We start by setting up your AI tools. Then how AI works, how it remembers and how to ask it well.

Up next

Lesson 3: Getting the most out of your AI tools

Set up your AI tools, link them to your notes, see how they work and ask them well.

Wednesday 11 November, 08:40 and 10:20

Open the Lesson 3 slides

← Lesson

1 of 27