EKALAVYA ACADEMY | SHORT BOOKS

Your AI rules on one page

What goes on a one-page AI rule sheet, how to draft it, how to bring people with you and how to keep it alive.

Ekalavya Academy by Almost Magic
Short Book 8 of 10 | Free to read and share

How to read this book

Before You Start

1. Learn
A few slides explain one idea in full sentences.

2. See it
A short, made-up example gives the idea a face.

3. Try it
A ten-minute exercise in the tools you already use.

4. Check it
Three questions. Guidance is at the back, not next to the question.

5. Keep it
A four-sentence summary to come back to.

Two levels in every chapter. The main slides are written so a first-time reader can follow them. A "Going deeper" slide gives the same idea at the level an experienced reader would argue about.

This book works with any AI tool you already have. It never asks you to buy or sign up for a particular product.

Where this book sits in the series

The Plan

Book 1	What AI is, and what it is not	Foundations
Book 2	Productivity with AI: writing, summaries, research, meetings	Practical use
Book 3	From one task to a repeatable workflow, and AI deputies	Workflow automation, agents
Book 4	Handling data safely	Data handling
Book 5	When a data shortcut becomes the rule	Data quality and governance
Book 6	Staying secure with AI	Security
Book 7	Rules, risk and what you must know	Risk and compliance awareness
Book 8	Your AI rules on one page	Policy and accountability
Book 9	When AI gets it wrong	Recovering from AI mistakes
Book 10	When not to use AI, and who stays in charge	Judgement and oversight

Each book works on its own. Read them in order or pick the one you need.

In this book

Contents

Chapter 1
What goes on a one-page rule sheet

Chapter 2
Drafting the sheet

Chapter 3
Bringing people with you

Chapter 4
Keeping it alive

At the back
Guidance and answers for every question and exercise

What goes on a one-page rule sheet

Chapter 1

By the end of this chapter you will be able to:

Say why a short rule sheet beats a long policy.

List the parts of a one-page sheet.

Decide what it must cover for your business.

Know what to leave out.

Time: about 40 minutes, including the exercise.

Why one page

Learn

A long policy is rarely read. A one-page sheet gets read, remembered and used.

The National AI Centre suggests a senior person own AI use and that an AI policy tell staff about AI risks and how to manage them. A one-page sheet is a small business version of that.

If a rule does not fit on the page, it is probably too complicated to follow.

IN PLAIN WORDS

If people will not read it, it is not a rule.

The parts of the sheet

Learn

Purpose
One sentence on why the sheet exists.

Allowed uses
What staff may use AI for.

Not allowed
What is off limits, and why.

Data rules
What may and may not be shared, using your labels.

Who approves
The named person for each kind of decision.

How to report
Who to tell when something goes wrong.

Rules people can follow

Learn

Write each rule so a new starter could follow it on day one. Use plain words and give one example.

Say what to do, not only what to avoid. "Check every client number against the source" beats "be careful".

Use the same words as the rest of your business, including your data labels.

RULE OF THUMB

Specific, short and checkable.

What to leave out

Learn

Tool lists that go stale
Say which kinds of use are allowed. Keep a separate register for tools.

Legal language
Point to an adviser for legal questions. Do not try to write law.

Everything that might happen
Cover the main cases and say who decides the rest.

Rules you will not enforce
An ignored rule teaches people to ignore the others.

Rules that hold up

Going Deeper

Principles plus examples
A principle gives direction. An example shows the line. You need both.

Decide the grey areas
Name the person who answers "can I use AI for this?" so people ask, not guess.

Match the risk
Strict where people or sensitive data are affected, light for low-risk drafting.

Make it findable
A sheet in a drawer does not help. Put it where people work.

A short example

See It

Kofi runs a ten-person design studio. His first AI policy ran to six pages and nobody read it.

He cut it to one page: a purpose line, five allowed uses, four things not allowed, three data rules, one approver and a reporting address.

In the next team meeting, three people quote it without looking.

THE LESSON

Short enough to remember is short enough to follow.

This is a made-up example for teaching.

Sketch your sheet

Try It

1. Write the six headings on one page.

2. Fill each with one or two lines, in plain words.

3. Cross out anything you will not enforce.

4. Read it aloud. Cut anything that stumbles.

YOUR TOOLS

Use paper or any notes tool. This is a first draft.

GUIDANCE

What to look for is in the Guidance section at the end.

Check yourself

Check It

Question 1
Why is a one-page sheet better than a long policy?

Question 2
Name four of the six parts of the sheet.

Question 3
Why leave out rules you will not enforce?

Answer in your own words before you read the Guidance section at the end.

Chapter 1 in four sentences

Keep It

KEEP THIS

A short rule sheet gets read and used. It has a purpose, allowed uses, things not allowed, data rules, who approves and how to report. Write each rule so a new starter can follow it, say what to do as well as what to avoid and leave out what you will not enforce.

Drafting the sheet

Chapter 2

By the end of this chapter you will be able to:

Write the purpose and the allowed uses.

Set clear limits and data rules.

Name who approves what.

Write a simple way to report problems.

Time: about 40 minutes, including the exercise.

Purpose and allowed uses

Learn

Purpose
"We use AI to save time on drafting and sorting, and a person is always responsible for the result."

Allowed
Drafting, summarising, brainstorming and tidying with no sensitive data.

Allowed with a check
Anything a customer will see, and any figures, with a person checking against the source.

Say who
Who may use which tools, if it differs.

Limits and data rules

Learn

Never share
Customer personal data, confidential files and passwords with a public tool.

Not for decisions
Decisions about people, such as hiring or firing, are made by a person.

Say when AI is involved
Tell customers where AI helps and where it could be mistaken for a human.

Use your labels
Link to the data labels from Book 4 so people know what counts as sensitive.

Who approves what

Learn

Only a person approves. Name the person, not just the role, for each kind of decision.

For example: the owner approves new tools and new kinds of use. The team lead approves sending any AI-assisted work to a customer.

If the person is away, name who covers.

REMEMBER

Only a human approves.

How to report a problem

Learn

Who to tell
One named person or address.

What to include
What happened, what tool and what data was involved.

No blame
Say that reporting quickly is what matters, and that honest reports are welcome.

What happens next
Say how the report is handled and when people will hear back.

Writing for real life

Going Deeper

Test it on a real case
Run last week's work through the sheet. Where does it say yes, no or nothing?

Make the exceptions explicit
If some uses need a check, say how and by whom.

Mind the supply chain
If a supplier uses AI on your work, your sheet should say how you ask about it.

Keep sensitive data in mind
Use the same labels and boundaries as in Books 4 and 5.

A short example

See It

Lena runs a small legal support service. She drafts her sheet and tests it on last week's work.

The test shows a gap: staff were summarising client letters with a public tool. She adds a line, "Client documents never go into a public tool", and names the approved alternative.

The sheet now covers a real case, not a guess.

THE LESSON

A test on real work finds the gaps.

This is a made-up example for teaching.

Draft the six parts

Try It

1. Write the purpose in one sentence.

2. List the allowed uses and the limits.

3. Name the approver for each kind of decision.

4. Test it on one real task from last week.

YOUR TOOLS

Use paper or any notes tool. Do not paste real client data.

GUIDANCE

What to look for is in the Guidance section at the end.

Check yourself

Check It

Question 1
Why name a person, not just a role, as the approver?

Question 2
What belongs in a report about a problem?

Question 3
Why test the sheet on a real case?

Answer in your own words before you read the Guidance section at the end.

Chapter 2 in four sentences

Keep It

KEEP THIS

Write a one-sentence purpose, a list of allowed uses and limits, data rules in your own labels and a named approver for each decision. Make reporting easy and blame free. Test the draft on a real task to find the gaps.

Bringing people with you

Chapter 3

By the end of this chapter you will be able to:

Introduce the sheet in a way people accept.

Train for use, not just awareness.

Handle questions and pushback.

Make it easy to ask.

Time: about 40 minutes, including the exercise.

Explain the why

Learn

People follow rules they understand. Start with what could go wrong and what the sheet protects: customers, the business and themselves.

Invite questions. A rule that comes with a reason is easier to follow.

Show a real example of the sheet in use.

IN PLAIN WORDS

Explain the why before the rule.

A short training session

Learn

Ten minutes
Walk through the sheet, one example per part.

One exercise
Everyone runs one real task through it.

One question each
Ask what is unclear or missing.

One owner
Say who to ask afterwards.

Common pushback

Learn

"This slows me down"
Show the low-risk uses that stay quick, and the few that need a check.

"Everyone does it"
Explain the risk to the customer and to them, without blame.

"I did not know"
That is the point of the sheet. Make it easy to find and ask.

"What about my tool?"
Have a simple way to ask for a new tool to be approved.

Make it easy to ask

Learn

People break rules when asking is hard. Name one person or channel for "can I use AI for this?" and answer quickly.

Share the answers, so others learn from them. Turn repeats into new lines on the sheet.

The National AI Centre guidance also suggests training people who oversee AI so they know when to step in.

RULE OF THUMB

A quick yes or no beats a quiet workaround.

Culture over compliance

Going Deeper

Reward good reports
Thank people who raise a problem early.

Lead by example
If the owner skips the sheet, nobody follows it.

Fit the people
A new starter, a manager and a contractor need different explanations.

Check understanding
Ask people to explain a rule in their own words.

A short example

See It

Dev owns a small online shop with six staff. He runs a ten-minute session and asks each person to run one real task through the sheet.

Two people find a gap: they use AI to write customer replies and are unsure what to check. He adds a line and names who to ask.

A week later, someone reports a mistake early because it is easy to do.

THE LESSON

Easy to ask means early to report.

This is a made-up example for teaching.

Plan your ten-minute session

Try It

1. Pick the three examples you will show.

2. Choose one real task for the exercise.

3. Name who answers questions afterwards.

4. Book the session in the calendar.

YOUR TOOLS

Use paper or any notes tool.

GUIDANCE

What to look for is in the Guidance section at the end.

Check yourself

Check It

Question 1
Why explain the reason before the rule?

Question 2
How can you handle "this slows me down"?

Question 3
Why name one person for "can I use AI for this?"

Answer in your own words before you read the Guidance section at the end.

Chapter 3 in four sentences

Keep It

KEEP THIS

Explain the why, run a short session with a real exercise and make asking easy. Expect pushback and answer it with examples. A quick yes or no beats a quiet workaround, and a thank-you for early reports keeps problems small.

Keeping it alive

Chapter 4

By the end of this chapter you will be able to:

Review the sheet on a schedule.

Learn from incidents and near misses.

Keep a version log.

Keep evidence that your people understand AI use.

Time: about 40 minutes, including the exercise.

Review on a schedule

Learn

Put a review in the calendar, for example every quarter. Ask what changed: new tools, new uses, new rules and new problems.

Review straight away after any incident or major change.

Check that the register and the sheet still match.

REMEMBER

A sheet that is not reviewed goes out of date.

Learning from incidents

Learn

Record
What happened, what was affected and who noticed.

Fix
Put it right for the people affected.

Cause
Why did the rule not stop it?

Change
Update the sheet or the training, and say what changed.

A simple version log

Learn

Version and date
For example v1.1, 12 January.

What changed
One line in plain words.

Who approved
The named person.

Who was told
How people found out about the change.

Evidence of AI literacy

Learn

The EU AI Act Article 4 asks organisations in scope to support their people's AI literacy, and does not require a guaranteed level for any individual.

Whether or not it applies to you, simple evidence helps: dates of sessions, who attended and what was covered.

Ask an adviser whether the Act applies to you.

RULE OF THUMB

Show what you did, not only what you meant.

Governance that scales

Going Deeper

Grow with your use
As use grows, the guidance moves from policy to a person per system, then to testing and monitoring.

Near misses count
A mistake caught before harm is the cheapest lesson you will get.

Avoid review fatigue
Keep it short. A fifteen-minute review that happens beats a long one that does not.

Look outside
Check regulators and guidance for changes at each review.

A short example

See It

Priya runs a small marketing agency. Every quarter she reviews her one-page sheet for fifteen minutes.

This quarter, two staff started using an AI note-taker. The review catches that it records client calls, and she adds a rule about telling clients first.

She logs the change as version 1.2 and tells the team on Monday.

THE LESSON

A short regular review catches what changed.

This is a made-up example for teaching.

Set up the loop

Try It

1. Put a quarterly review in your calendar.

2. Create a version log with the four columns.

3. Write down your first training session date.

4. Keep an incident note template.

YOUR TOOLS

A table in any tool, or on paper.

GUIDANCE

What to look for is in the Guidance section at the end.

Check yourself

Check It

Question 1
What should a review look at?

Question 2
What four things belong in a version log?

Question 3
Why keep simple evidence of training?

Answer in your own words before you read the Guidance section at the end.

Chapter 4 in four sentences

Keep It

KEEP THIS

Review the sheet on a schedule and after any incident. Learn from incidents by recording, fixing, finding the cause and changing the sheet. Keep a version log and simple evidence of training. Ask an adviser whether rules such as EU AI Act Article 4 apply to you.

Guidance and answers

At The Back

Questions and exercises stay in the chapters, on their own slides. Guidance lives here, so you can try first and look afterwards.

Try the task or question in your own words, then compare. Where your answer differs, that is worth a note, not a correction.

WHERE TO FIND IT

The Guidance section below: answers and exercise guidance, one part for each chapter.

Chapter 1: answers and guidance

Guidance

Answer 1
People read, remember and use it. A long policy is rarely read.

Answer 2
Any four of: purpose, allowed uses, not allowed, data rules, who approves, how to report.

Answer 3
An ignored rule teaches people to ignore the others.

Exercise
A good draft has no sentence you could not say out loud. If a line is vague, add an example.

Chapter 2: answers and guidance

Guidance

Answer 1
A named person is clear and cannot be passed around. Say who covers when they are away.

Answer 2
What happened, which tool and what data was involved.

Answer 3
It shows where the sheet says yes, no or nothing, so you can fix the gaps.

Exercise
A good test finds at least one gap. Fix it, and test again with a different task.

Chapter 3: answers and guidance

Guidance

Answer 1
People follow rules they understand, and a reason makes the rule easier to remember.

Answer 2
Show that low-risk uses stay quick and only a few need a check.

Answer 3
It makes asking easy, so people ask instead of guessing or working around the rule.

Exercise
A good plan has a real task and a named person. Run it within a week while the sheet is fresh.

Chapter 4: answers and guidance

Guidance

Answer 1
What changed: new tools, uses, rules and problems, and whether the sheet and register still match.

Answer 2
Version and date, what changed, who approved and who was told.

Answer 3
It shows what you did, and helps if a customer or regulator asks.

Exercise
A good setup has dates, owners and a log. If any is missing, add it before you close this book.

About Ekalavya Academy

Find Out More

Ekalavya Academy is a learning series from Almost Magic Tech Lab. The books are written for people who teach themselves, using the tools they already have.

Sources used: National AI Centre, "Guidance for AI adoption: foundations" (ai.gov.au), and the European Commission AI Act Service Desk, Article 4 (ai-act-service-desk.ec.europa.eu). Awareness, not legal advice.

Almost Magic builds tools that help people check and govern their AI use. The books never depend on those tools.

LINKS

Almost Magic: almostmagic.net.au

Case studies: ai-casestudies.almostmagic.net.au

Page for this book on the Academy site: link to be added when the page exists.

Ekalavya Academy by Almost Magic | almostmagic.net.au