Artificial Idea | AI careers · practical prompts · no hype Thursday, August 28, 2025 · Issue #8 · Prompt Tutorial
The best marketers are not working harder than everyone else. They are working with better tools. Here is the stack.
Monday's breakdown of the WEF report prompted more direct replies than any issue we have published. Most of them came from managers and senior professionals asking a version of the same question: this is useful context, but what do I actually do with it on Monday morning?
Thursday issues exist to answer that question. And this week, we are answering it for one of the most time-pressured professional categories in any organisation: marketing.
Marketing teams carry an unusual burden in the AI transition. They are expected to produce more content, across more channels, for more fragmented audiences, with the same or smaller headcount than two years ago. The expectation that AI solves this problem is widespread. The practical knowledge of how to make it do so is considerably less common.
These five prompts are the ones that actually move the needle. Not in theory. In the specific, recurring tasks that consume a marketer's week before they get to the work that actually requires their judgment.
Prompt 1: The campaign brief builder
The problem it solves: writing a campaign brief that is specific enough to be useful, aligned across stakeholders, and clear enough that a creative team can actually execute against it without three rounds of clarification.
You are a senior marketing strategist with deep experience
briefing creative teams on integrated campaigns.
Context: [describe your product or service, your target audience,
the campaign objective, the key message you want to land,
and any constraints such as budget, timeline, or channels]
Task: Write a complete campaign brief. Include: campaign objective
in one sentence, target audience profile, core message,
supporting messages (maximum three), tone of voice guidance,
channel recommendations with rationale, and success metrics.
Constraints: Keep each section concise. Use plain language.
Flag any areas where you need more information from me
before the brief can be finalised. No marketing jargon.
Where most campaign briefs fail is in the gap between what the brief says and what the creative team understood it to say. The instruction to flag information gaps before finalising forces the model to surface those gaps explicitly, which means you catch misalignments at the brief stage rather than after the first creative review.
Prompt 2: The audience insight extractor
The problem it solves: turning a pile of customer feedback, survey responses, or review data into actionable insight without spending half a day reading everything manually.
You are a consumer insights analyst. Your job is to find the
patterns that matter, not to summarise everything equally.
I am going to paste a set of customer responses. Please do
the following:
1. Identify the three most common underlying concerns or desires,
stated in the customer's own language where possible
2. Identify any tension or contradiction in what customers say
they want versus what they seem to actually value
3. Pull out the two or three most quotable verbatim lines that
best represent the overall sentiment
4. Tell me the one thing this data suggests we are currently
getting wrong
Here is the data: [paste responses]
Point four is the one most people leave out. It is also the one that produces the most useful output, because it forces the model toward a diagnostic conclusion rather than a neutral summary. Neutral summaries are rarely what you need.
Prompt 3: The content repurposer
The problem it solves: getting more distribution from content you have already produced, without starting from scratch for every channel.
You are a content strategist who specialises in adapting
long-form content for multiple channels without losing
the core argument.
I have written a [blog post / report / interview / webinar
transcript] on the following topic: [topic]
Here is the original content: [paste content]
Please produce the following from this single piece:
1. A LinkedIn post (150 words, professional tone, ends with
a question to drive comments)
2. Three tweet-length takes (under 240 characters each,
each making a distinct point from the original)
3. A five-bullet email summary for our newsletter audience
(each bullet one sentence, leads with the insight
not the topic)
4. One follow-up content idea this piece naturally sets up
Constraints: Do not add claims that are not in the original.
Maintain the tone and positioning of the source material.
Flag any section where the original is too thin to
repurpose confidently.
The constraint at the end matters more than it appears. Without it, AI will confidently fill gaps in your source material with plausible-sounding content that you did not write and may not be able to stand behind. Explicit instruction to flag thin sections gives you control over what gets published.
Prompt 4: The competitive positioning analyser
The problem it solves: understanding quickly how competitors are positioning themselves and where the gaps in the market are, without commissioning a full competitive analysis.
You are a brand strategist with expertise in competitive
positioning. You are sharp, direct, and not interested
in telling me what I want to hear.
I am going to give you the homepage copy or key messaging
from three of our competitors. After reviewing them, please:
1. Identify the positioning territory each competitor
is claiming
2. Identify any positioning territory none of them
are claiming that represents a genuine opportunity
3. Tell me the one word or phrase each competitor
owns in the market right now
4. Give me your honest assessment of where our
current positioning is weakest relative to this
competitive set
Here is the competitor messaging: [paste copy]
Here is our current positioning: [paste your messaging]
The instruction to identify unclaimed territory is the most valuable part of this prompt. Most competitive analysis tells you what competitors are doing. This one is designed to tell you what they are not doing, which is where differentiated positioning lives.
Prompt 5: The performance debrief
The problem it solves: turning a campaign postmortem into something more useful than a list of what happened, by extracting the learning that should change what you do next time.
You are a senior marketing director running a campaign
debrief. You are more interested in extracting learning
than in assigning credit or blame. You ask uncomfortable
questions.
Context: [describe the campaign, its objectives,
what happened, and the key metrics]
Please structure the debrief as follows:
1. Where the results matched expectations and why
(do not dwell here)
2. Where the results diverged from expectations and
the most likely explanations
3. The assumption we made that turned out to be wrong
4. The one thing we should do differently next time,
stated as a specific operational change not a
general principle
5. The question this campaign leaves unanswered that
is worth investigating before the next one
Constraints: Be direct. Do not soften the diagnosis.
If the data I have given you is insufficient to draw
a conclusion, say so rather than speculating.
Point three is the one that separates useful postmortems from ones that feel thorough but produce no change. Identifying the specific assumption that was wrong is considerably more actionable than identifying the metric that missed target. The metric tells you what happened. The assumption tells you why, and what to think differently next time.
How to use this stack
Do not try all five this week. Pick the one that matches the most painful recurring task in your current workload, run it once with real content, and spend time editing the output rather than rewriting it from scratch. The editing to rewriting ratio is a reliable signal of whether the prompt is working. If you are rewriting, the prompt needs more context or tighter constraints. If you are editing, it is working.
The goal is not perfect output on the first run. It is output that is good enough to edit in less time than it would take to write from zero. That threshold, applied consistently across five recurring tasks, is where the three hours a week come from.
Monday we are covering something that sits underneath every piece of career advice about AI and has not been addressed directly yet: the specific skill that makes someone genuinely hard to replace, regardless of their profession or seniority level. It is not the skill most people are currently investing in.
See you then.
— The Artificial Idea team
