A Guide to Prompting with LLMs: 2025 Edition

There are many prompting frameworks available today, most share the same principles. Try a few and choose the one that fits you best.

Before crafting a prompt, think about four components. This simple checklist helps you structure the request so the AI can respond appropriately.

ChatGPT Framework: GRACE

G – Goal: What you want

R – Role: Who the model should act as

A – Assets: What inputs you’re giving it

C – Constraints: Style, length, tone, limits

E – Expectation: What “good” looks like

Microsoft Framework: Goal + Context + Source + Expectations

Google Framework: Persona + Task + Context + Format

Anthropic Framework: Role → Context → Instructions → Constraints → Revision

Perplexity Framework: Question → Context → Action → Format

The following techniques are essential for reliable AI outputs. Each tip includes a marketing‑related example with a bad prompt and a good prompt to illustrate how clarity and context change the result.

Be Clear and Specific About the Goal:

A clear task description and desired output will cut down unnecessary back-and-forth.

Example: Campaign planning – Channel role map

Example: Draft a creative brief for a fitness app campaign targeting women 25-34; include sections for objective, audience insight, tone, key message, and KPIs.

Provide Context and Data:

LLMs perform better when given relevant background information. Include why you’re doing the task and any important facts. When the task involves numbers (e.g., campaign performance), include the data you want analysed.

Example: Reporting – ROAS truth vs platform reporting

Assign a role:

Telling the AI to adopt a persona helps guide its style (formal, friendly, creative, etc.), and depth of expertise. In marketing, that role could be “product manager”, “data strategist”, “programmatic buyer” or “creative director”.

Example: Audience creation – Define a target segment

Example: You are a media planning expert. Explain these social media metrics in simple terms for a client.

Specify format and constraints:

Formatting instructions make outputs easier to use such as tone, length and structure.

You can ask for [5] sentences, [500] words, [5] bullet points. You can also request a table and list the columns you need (or refine them together with the model), or JSON if needed. Some tools also let you generate slides that can be copied straight into presentations.

Example: Creative asset production – Content calendar

Example: List 5 audience segments as bullet points with a one-line description each.

Double‑check the facts and respect limitations:

Large language models can occasionally generate convincing but inaccurate information. To reduce the risk:

• Check credibility. Ask the model to cite its sources, then verify those references against reliable data.
• Protect confidentiality. Don’t include proprietary or sensitive material unless you have explicit approval to do so.
• Stay concise. Overly long prompts may be cut off or misinterpreted; focus on clarity over length.
• Iterate deliberately. Begin with a focused prompt, review the response, and refine your request step by step until the output meets your needs.

These approaches aren’t mandatory, but they can elevate the quality, relevance, and consistency of AI-generated work.

Treat the AI as a collaborator: Use an iterative, multi-turn process. Think of the AI as a creative partner rather than a one-shot tool. Start with a base prompt, review the response, then refine it — adjusting tone, adding missing details, or asking for new angles.

Example: Good start — now make it more concise and include three alternative taglines.

Define rules and guardrails: If there are words, topics, or claims the output should avoid, make that clear in your prompt. Setting these boundaries early helps maintain brand safety and alignment.

Example: Draft a social caption for our energy drink launch. Avoid any references to alcohol, nightlife, or extreme sports. Keep the tone confident but inclusive.

Provide examples or inspiration: Show the AI what “good” looks like. Share snippets of past campaigns, preferred tone of voice, or sample copy to steer the output toward your desired style.

Example: Here’s an example of the tone we use: ‘Smarter energy, designed for everyday athletes.’ Now write three alternative headlines for our new flavour launch using the same voice.

Let the AI interview you: Encourage the model to ask clarifying questions before it answers. This helps it fill information gaps / blind spots and tailor the response more precisely to your needs.

Example: Before suggesting audience segments or messaging, ask me up to three questions to clarify who my target customers are — for example, their demographics, behaviours, or purchase motivations. Once you have my answers, propose tailored audience personas with key insights.

Use reusable variables: Write flexible prompts using placeholders like {brand}, {audience}, or {budget}. This lets you repurpose and scale prompts across different clients or campaigns without rewriting them each time.

Example template: Write a {length}-word product description for {brand}, targeting {audience}. Highlight how {product feature} helps solve {customer problem}.

Filled Example: Write a 100-word product description for EcoHome, targeting urban renters. Highlight how SmartTherm helps reduce energy bills without compromising comfort.

Build your own prompt library: Save the versions that consistently deliver strong results. A small, curated set of tested prompts becomes a powerful reference for future projects and training new team members.

How:

Store directly in the AI platform (process differ by platforms): pin the chat, rename it, search bar; keep one pinned chat called Prompt Library.

Or store out of AI-platform like a word document or spreadsheet (category, prompt, use case as potential columns), Notion… – make them searchable, shareable with your teams