1.4 Audience-Tailored Prompting

Tip

Audience-tailored prompting means telling the AI who the output is for. By naming the reader, listener, or user, the AI automatically adjusts vocabulary, complexity, tone, and structure to match.

When to use it

Cross-expertise communication

You need to explain something to an audience whose background differs from yours: a funder, a journalist, a student, or a colleague in another discipline.

Science communication

You are translating research findings for the public, a policy brief, a grant impact section, or teaching materials.

---

What changes when you specify an audience

The same question produces fundamentally different output depending on who you say it is for:

Audience	Vocabulary	Depth	Structure	Tone
Fellow specialist	Technical jargon, no definitions	Deep, assumes prior knowledge	Dense paragraphs, citations	Formal, precise
Researcher in another field	Some jargon, briefly defined	Moderate, explains key concepts	Clear sections, signposted	Professional, accessible
Policy maker	No jargon	High-level, action-oriented	Bullet points, executive summary	Direct, impact-focused
General public	Everyday language, analogies	Broad strokes only	Short paragraphs, narrative	Warm, engaging
Student	Simplified terms, definitions	Step-by-step	Examples, practice questions	Encouraging, patient

Specifying the audience is often the single most efficient way to control all of these dimensions at once.

---

Best practices

1. Be specific about who the reader is - “A soil ecologist reviewing for Journal of Ecology” works much better than “a scientist.” The more specific, the better the AI can tailor its output.

2. State what the audience already knows - “Assume the reader is familiar with mixed-effects models but not with Bayesian methods.” This prevents over-explaining things your reader knows while ensuring new concepts are covered.

3. Combine audience with format and tone - Audience alone goes a long way, but pairing it with format (“bullet points”, “narrative prose”) and tone (“formal”, “conversational”) gives you precise control.

4. Adjust format to match audience expectations - Policy makers expect short summaries with action items. Peer reviewers expect structured sections with citations. Do not change the audience without changing the format.

5. Watch for accuracy loss when simplifying - When asking the AI to simplify for non-experts, add: “Simplify language but preserve scientific accuracy. Do not omit caveats or limitations.” Without this, the AI may strip important nuance alongside jargon.

---

Ready-to-use prompts

Expert to public

Translate research findings for a general audience:

Prompt: Research for the public

Rewrite the following research abstract for a general audience
with no scientific background. Use everyday language, short
sentences, and at least one analogy. Preserve the core findings
accurately and do not omit caveats or limitations.

Abstract:
[PASTE YOUR ABSTRACT HERE]

Note

Adding “do not omit caveats” prevents the AI from oversimplifying in ways that misrepresent your findings.

Multi-audience

Generate the same content at multiple levels in one go:

Prompt: Three-level explanation

Explain the concept of trophic cascades in three versions:

1. For a fellow ecologist (use technical terminology freely,
   focus on recent debates and mechanisms)
2. For a master's student in environmental science (define
   key terms, include a concrete example from a well-known
   study)
3. For a Dutch provincial policy maker (no jargon, focus on
   why this matters for nature management in the Netherlands,
   keep it under 150 words)

Note

This multi-pass approach is useful when you need the same content for different stakeholders, for example a research summary, a teaching slide, and a policy brief.

Grant impact section

Write a societal impact paragraph tailored to a funding panel:

Prompt: Impact section for NWO

I am writing the societal impact section of an NWO Open
Competition proposal on [YOUR TOPIC].

The audience is the NWO assessment panel: experienced
researchers who value concrete, measurable impact over vague
promises. They are familiar with Dutch science policy
(Nationale Wetenschapsagenda, Nationaal Programma Landelijk
Gebied) but not necessarily with my specific subfield.

Based on the following research summary, draft a 300-word
impact section that connects my research to tangible societal
benefits. Be specific about who benefits, how, and on what
timescale.

Research summary:
[PASTE YOUR SUMMARY HERE]

Note

Notice how this prompt specifies what the audience knows (“familiar with Dutch science policy”) and does not know (“not necessarily with my specific subfield”). This is one of the most effective things you can do.

---

The spectrum from simple to expert

A useful mental model is the complexity spectrum. You can place your audience anywhere along it:

Prompt: Set complexity level

Explain [TOPIC]. Target the explanation at the level of
[CHOOSE ONE]:
- a curious 10-year-old (simple analogies, no jargon)
- a university student encountering this for the first time
- a researcher in a related but different field
- a specialist who wants nuance and edge cases

You can also ask the AI to shift along this spectrum after the fact: “Now rewrite that explanation for a policy audience” or “Make this more technical for peer review.”

---

When not to use it

• Factual lookups - “What is the boiling point of water?” does not benefit from audience specification.
• Code generation - Unless documentation is involved, the task itself defines the audience.
• When the task already implies the audience - If there is no ambiguity about who will read the output, do not over-specify.

A good test: if a colleague reading your prompt would ask “Who is this for?”, you need to specify the audience. If the answer is obvious from context, skip it.

---

Key takeaway

Remember

Naming your audience is often the single most efficient prompt improvement you can make. One specific sentence about who the reader is, what they know, and what they need can replace a paragraph of instructions about tone, vocabulary, and depth.

---

Sources

Based on materials from Vanderbilt Prompt Pattern Catalog, COSTAR Framework, Bsharat et al. (2023) - 26 Principles, and Trott (2024) - Measuring Readability.

Have audience-tailored prompts that work well for your research communication? Share them with RSO so we can add them to the library.