Role, Marketing

Claude for
marketing

Copy, campaigns, landing pages, content repurposing, and the brand voice that keeps everything feeling like one company.

Marketing is where Claude misuse is most visible and also where the best usage hides in plain sight. The visible misuse is generic AI copy on landing pages and subject lines that sound like every other SaaS company. The good usage is invisible because it happens in the drafting phase, and a real marketer edits it into something that sounds like them. This page is about the good usage.

Three places to start

Three workflows where Claude amplifies a good marketer without flattening their voice.

1. Iterating landing page copy

One version of a headline to argue about is a bad meeting. Five versions with different angles is a productive one. Ask Claude for five rewrites of your headline and subheadline, each with a different approach (problem-first, outcome-first, social proof, specific example, contrarian), and an explanation of what each one is doing and who it's likely to resonate with. You're no longer debating whether the copy is good. You're debating which kind of good it should be.

Copy iteration prompt, annotated

Five rewrites in a single prompt. Four elements turn this from "another AI copy dump" into something worth reading in a meeting.

YouHere is the current headline and subheadline for our cost-tracking feature landing page, pasted below. Here is the brief it was written for, pasted below. 1 Here is our brand voice guide, pasted below. It is deliberately specific: direct, no hedging, no "unlock your potential" language. 2 Produce five rewrites of the headline and subheadline. Each should take a different angle: problem-first, outcome-first, social proof, specific example, contrarian. For each, one line on what it is doing differently and who it is most likely to resonate with. 3 Do not use any phrase that would feel at home in a SaaS landing page template. If a rewrite drifts into generic copy, flag it yourself rather than including it. 4
1
Context. The feature, the current copy, the brief. Grounds every rewrite in the same target.
2
Example. Voice guide with explicit "not this" language. Sharper than "match our brand voice".
3
The ask. Five specific angles plus a one-line rationale each. Turns the output into a decision, not a menu.
4
Escape hatch. Tells Claude to self-police against generic AI copy. Surprisingly effective.

Where this falls over

If you don't give it a proper voice reference, the output will sound like generic SaaS copy. Always include at least one example of copy you were happy with, and ask it to match that voice.

2. Repurposing one piece across channels

Long-form content (a blog post, a whitepaper, a webinar recap) is expensive to produce and usually lives on one page. Claude can repurpose it into a LinkedIn post, a Twitter thread, a newsletter intro, a Slack summary, and three quote-style image cards in one prompt. The repurposing is usually 80% of the work the content team does manually.

One blog post into five channels, weak vs strong

The difference between "Claude wrote this" and "Claude helped me write this".

Yourepurpose this blog post for social
What Claude sends back

A LinkedIn post that starts "Exciting news!" and ends with "Thoughts?". A Twitter thread where every tweet begins with a number emoji. Three "quote cards" with phrases nobody would actually quote. Generic marketing voice, reads as AI from the first word.

YouHere is a 1,200-word blog post about our Q1 customer research, pasted below. Here is our brand voice guide, pasted below. No "exciting news", no "thoughts?", no emojis, no engagement-bait. Produce versions for: a LinkedIn post (under 1,300 characters, first-person voice, no call to action), a short Twitter thread (six posts max), a newsletter intro paragraph (150 words that make people want to click), a Slack summary for our internal #marketing channel (five bullets), and three pull quotes I could turn into image cards (each under 20 words, each must be something a real person would actually quote). Match the voice of the original post. Do not invent any quote or statistic that is not in the source.
What Claude sends back

Five versions, each matching the voice of your original post. The LinkedIn post opens with a specific finding from the research. The Twitter thread is six posts, none of them starting with a number. The newsletter intro is 148 words and ends with a proper hook, not a CTA. The Slack summary is five bullets, bluntly written. The three pull quotes are all direct language from the source, and each one is something a reader could actually imagine saying out loud.

Where this falls over

Without tight voice constraints, the repurposed versions will drift into generic marketing speak. The fix is the same as always: paste examples of the voice you want.

3. Turning strategy into a campaign brief

Strategy ends in a paragraph like "we should run a campaign targeting CFOs about cost predictability". That paragraph is not a brief. Claude will turn it into one: objective, success metric, audience, core message, channels, content list, and crucially, the top three decisions that need to be made before the team can start. The last part is where a good brief separates itself from a bad one.

LK
Liz K. #marketing Today, 11:20

Quick share. Took the strategy paragraph from last week's leadership review and asked Claude to turn it into a campaign brief. Normal structure, but the critical line was "flag any decisions that still need to be made rather than filling them in".

The brief came back 90% written. The remaining 10% was three decisions flagged as "need to be answered before we start": budget, which persona we are targeting first, and whether we are aiming for pipeline or brand. Those three questions are now the agenda for next Tuesday's meeting. An hour saved and the meeting is about decisions instead of a readout.

7 clap 3 heart

Where this falls over

Claude will paper over the gaps in your strategy by inventing plausible details. Explicitly ask it to flag anything ambiguous rather than filling it in.

Worked examples

Before and after, on tasks from marketing work.

Prompts to steal

Copy any of these, fill in the square brackets, run.

Try this week

Pick one. All of these are under an hour.

The real test of Claude in marketing is whether the output still sounds like your brand after three rounds of iteration. If it does, the tool is working. If it drifts back to generic by round three, your voice reference isn't strong enough.

If you want to go broader, the Beginner guide covers the habits that apply everywhere. The prompt library has the full set of role-tagged prompts in one place.