Prompting for Marketers — Kerry Cokelekoglu

Generic Prompts Get Generic Output

"Write me a blog post about email marketing" will get you a blog post about email marketing — the kind that already exists 10,000 times over. The gap between people who find AI genuinely useful and people who call it "fine, I guess" almost never comes down to the tool. It comes down to the prompt.

Marketers are used to briefing people: an agency, a freelancer, a junior hire. A good brief includes who the work is for, what it needs to accomplish, and what "good" looks like. A prompt is a brief. Treat it like one and the output changes immediately.

For beginners

Before writing a prompt, write one sentence describing what a great human teammate would need to know to do this task well. That sentence is your prompt's missing context.

For experts

Save your highest-performing prompts as reusable templates with placeholder fields, not one-off chat messages. A prompt you'll use again is an asset, not a conversation.

The Framework: Role, Context, Task, Format

Every prompt that consistently performs well has four parts, in this order:

Role — who should the AI act as? "You are a senior lifecycle marketer" produces different output than no role at all.
Context — audience, offer, brand voice, channel, funnel stage, constraints. This is the part most people skip, and it's the part that matters most.
Task — the specific, concrete thing you need: not "help with email," but "write a 3-email win-back sequence for lapsed trial users."
Format — how the output should be structured: subject line + preview text + body, a table, a numbered list, a short executive summary.

Pro tip

Write the Format instruction last, but put it in the prompt early. Telling the AI the shape of the answer before it starts generating produces more structured output than asking for structure after the fact.

Add the Context That Actually Changes the Output

Not all context is equal. Telling the AI your company name rarely changes anything. Telling it the audience's biggest objection, the channel's character limit, or the one competitor you're positioned against will change the output every time. When you're deciding what to include, ask: "If I removed this sentence, would the answer actually be different?" If not, cut it.

Iterate Instead of Restarting

The biggest efficiency gain in prompting isn't writing the perfect first prompt — it's not throwing away an 80%-good answer and starting over. Tell the AI exactly what's wrong: "the tone is too formal for our audience," "cut this to half the length," "make the CTA more direct." Each round gets closer, and most outputs are three rounds away from great, not zero.

For beginners

Treat the first response as a draft, not a verdict. "Make the second paragraph punchier" is a faster path to a good email than rewriting your original prompt from scratch.

For experts

For complex tasks, ask for an outline first and approve the structure before generating full copy. Structural changes after a full draft cost more iteration than structural changes before one.

Build a Prompt Library, Not a Memory

If you find yourself typing a similar prompt for the fourth time, it should already be saved. A prompt library — even a simple shared doc — turns "I think I had a good prompt for this somewhere" into a 10-second copy-paste. Every recurring marketing task (weekly report, ad variant batch, email QA) deserves one home for its best prompt.

Pro tip

Name saved prompts by outcome, not by tool: "Weekly Performance Narrative," not "Claude Prompt 4." You'll find them faster, and so will your team.

The Takeaway

Prompting isn't a technical skill — it's a briefing skill marketers already have. The framework doesn't change what AI can do; it changes how clearly you ask for it. Role, context, task, format. Every time.

Interested in Learning More?

If you're a marketer building out your AI workflow and want to talk through how to make this work for your specific setup, reach out — I'd love to compare notes.

hello@kerrycokelekoglu.com kerrycokelekoglu.com