Lesson 4: Define the Audience Before You Ask for the Output

Lesson overview

When you do not tell AI who the output is for, it guesses. That guess is almost always too broad, too generic, and wrong for your actual situation. This lesson explains how to define audience clearly and why it changes almost everything about what AI produces.

What this means

Audience is not just a demographic detail. It shapes the vocabulary AI chooses, the depth of explanation it provides, the examples it reaches for, the level of formality it uses, and the assumptions it makes about what the reader already knows.

Two completely different pieces of writing can come from the same underlying request — just by changing who they are written for.

Why it matters

A technical explanation written for a senior engineer will be impenetrable to a first-time user. An explanation written for a first-time user will frustrate a senior engineer who already knows the basics.

Most default AI output lands somewhere in the middle — accessible to no one in particular and particularly useful to no one. Defining audience moves that needle immediately.

What most people do wrong

Skipping audience entirely

The most common version of this problem. The request mentions a topic but says nothing about who will consume it. AI defaults to a vague, generalized reader.

Using unhelpful audience definitions

"Our customers" tells AI almost nothing unless you add more. Who are your customers? What do they care about? What do they already understand? How do they talk?

Assuming AI knows the reader

Terms like "stakeholders," "the team," "technical users," or "executives" mean different things in different organizations. Tell AI what your version of those terms means.

Forgetting that tone is partly determined by audience

A direct, informal tone that works well for an internal Slack message will feel wrong in a board update. Audience should influence tone guidance too.

What better looks like

A useful audience definition includes:

Who they are — role, background, level of expertise
What they already know — what AI can assume vs. what needs explanation
What they care about — what problems they have, what outcomes they want
Where they will read this — the context in which the writing appears shapes expectations

You do not always need all four. But you should be able to answer them before you write the prompt.

Audience comparison: Same task, three audiences

Task: Explain what two-factor authentication (2FA) is.

For executive leadership:

Explain two-factor authentication in plain English for a non-technical executive audience. Focus on why it matters for company security and what implementing it broadly would mean for the organization. Assume no technical knowledge. Under 150 words.

Result: No acronyms, no implementation details, business impact framed around risk reduction.

For customers:

Explain two-factor authentication to a customer who has never set it up and may find security settings confusing. Keep it simple, reassuring, and focused on why it helps protect their account. Avoid technical jargon. Step-by-step encouragement rather than technical explanation.

Result: Plain language, reassuring tone, focused on the experience of setting it up.

For internal technical team:

Write a brief internal doc on 2FA for our engineering team covering TOTP vs. SMS-based authentication, current security tradeoffs, and our recommended implementation path. Assume the reader is a developer who knows authentication concepts.

Result: Assumes knowledge, goes deeper on tradeoffs, skips beginner explanations.

Same topic. Three very different outputs. Audience is the variable.

Weak example

Write a summary of our Q3 product updates.

What is missing: Who is reading this summary? A customer email and a board presentation and an internal team retrospective all have different audiences — and should read completely differently.

Strong example

Write a summary of our Q3 product updates for customers who receive our monthly newsletter. These are non-technical users who use our product to manage scheduling. Highlight two or three improvements that make the product easier to use, with brief one-sentence explanations of each. Tone is friendly and clear. Avoid internal product names or jargon. Under 250 words.

What is better: The reader's role, expertise level, content channel, and reading context are all defined. The output will be shaped accordingly.

Practical exercise

Take one of your most common recurring work tasks that involves writing. Rewrite the request three ways — once for each of these audiences:

Executive leadership or senior stakeholders — brief, outcome-focused, no technical detail
Customers or external readers — plain language, value-focused, empathetic
Internal teammates — direct, assumes shared context, practical

Compare the three outputs. Notice how AI adjusts depth, tone, and format based only on the audience change.

Reflection prompt

Who is the most common intended reader of the things you ask AI to write?
Have you ever sent AI output to someone and had it land wrong — too formal, too technical, too simple? What was the audience mismatch?
What is something specific about your audience that AI could not assume on its own?

Key takeaway

Audience shapes nearly everything about a piece of writing. Name the audience before you ask for the output — not just who they are, but what they know, what they care about, and where they are reading it.