AI-Driven GTM Imagery: Part 4, Pre-Gen Campaign Planning

There is a moment in every campaign where the upstream work is done and you feel ready to make things. You have your positioning. You have your brand system. You have your copy bank. You have your visual components. And you want to start producing ads. That pull is strong; but, resist it… for one more step.

In the pillar article for this series, I described the five-stage process for making production-quality marketing graphics with AI. The first three stages are about ingredients: strategy, brand identity, and copy. The fourth is about gathering the visual raw materials. The fifth is assembly. But I glossed over something critical between “having ingredients” and “assembling them,” and it’s the step that determines whether the final output feels like a cohesive campaign or a disconnected pile of ads.

That step is campaign planning. The creative brief. The shot list. The mood board. The omni-channel asset plan that tells you exactly what you’re making, for which channels, in what formats, with what messages, before anyone opens a design tool or writes a prompt. This is the work that separates campaigns that feel intentional from ones that feel like someone just started making things and hoped for the best.

The Brief That Launches Everything

Once I get to the “make the stuff” phase, I have all my core components stored in one place, one folder. And I am now asking the LLM to produce something. If this is a net new campaign that I haven’t made things or created for before, I start with a creative brief.

I really like Emily Kramer’s GACCS brief format. I turned that into a skill file, and I improve it over time as I find new things to add to the brief that help the LLM produce better output (get it at the end of the post, IYI). The brief covers all the basics. What are we advertising for? What business goal are we trying to achieve? To which target audiences? What are we offering? What are we trying to get them to do? What channels will it run in? What ad types are we running in each of those channels? How much are we spending? What does success for the campaign look like? How will we measure it?

That’s the brief you would give to your account team at an agency and say, “Alright, go make this campaign.”

The format matters less than the completeness. Whatever brief structure you use, it needs to answer every question a designer or media buyer would ask before they started working. Because the LLM is now playing both of those roles, and unlike a human creative team, it won’t push back and ask for what’s missing. It will just fill the gaps with generic assumptions that water down your entire campaign.

This is the “yes-man effect” that Superside’s 2025 research identified as a core design risk with AI: the model validates rather than challenges creative direction. Without clear constraints in the brief, AI output trends toward generic. A sharp brief is how you stay in control of the direction.

Getting to the 30,000-Foot View

After I have the campaign brief, I know my channels, my CTAs, my target customers, my offer, and I can start putting the creative together from a 30,000-foot view. This is where I do something that might surprise people who know how much I love Claude.

I take the campaign brief and put it into ChatGPT. I find that ChatGPT is a very compliant model, and it’s really good at doing work like this. What I mean by compliant is that it doesn’t take as much creative liberty as Claude often does. The thing I like about Claude is precisely that it takes creative liberty. It’s a brilliant brainstorming partner. But when you already have a pretty solid vision on a direction, I find ChatGPT is more useful because it’s not going to try and reinvent the wheel. It’s going to take whatever your vision is and execute it precisely, flawlessly, thoroughly.

I take the campaign brief, give it the context from the prior stages (the copy bank, the business positioning strategy, the product marketing strategy), and I’ll say:

“I want to design a campaign for the goals outlined in this brief. Based on the channels it’s going to run in and what the offer is, how would you use each channel to run this campaign? What would you produce in each place? How many variants would you produce in each place? And what role would each channel play?”

What it gives back is almost like an asset requirements list by channel and a bird’s-eye view of what the whole campaign looks like when it is put together. It will say “on Instagram you’re doing this, in email you’re doing this, on the website you’re doing this, in SMS you’re doing this”. It pulls all the channels from the brief, pulls the customer information and messaging from the background files, and gives you an omni-channel campaign asset plan.

This is the document that tells you what you’re actually building. Not at the pixel level, but at the architectural level. How many ads. What formats. Which messages map to which channels. What the customer journey looks like across touchpoints. It is the difference between “let’s make some ads” and “let’s execute a campaign.”

From Asset Plan to Shot List to Mood Board

The asset plan is the what. The next step is the how.

Once I have the omni-channel asset plan, I take it and create a shot list. A shot list is a very specific outlining of all the images I want to create that show how the campaign will come to life across all channels. If the asset plan says “three static ads for Instagram feed, two stories, one carousel,” the shot list describes each of those six assets in concrete visual terms. What’s the background? What’s the focal point? What text treatment? What mood? What format dimensions?

The shot list is where the brand system from earlier stages becomes essential. Every visual decision in the shot list references the design tokens, the color palette, the typography rules, the texture system. You’re not making aesthetic choices on the fly. You’re applying a system.

Next, I take that shot list and put it into something like Gamma to create a presentation that has a mood board aesthetic. It shows all the different assets the campaign is comprised of, how they look in each channel, the types of CTAs they might have, the type of overall vibe. The goal is to show the team and the client what the campaign will look like when it comes to life across all the different places it runs.

That gives you a stop point. Do I like this? Do I want the mood to be a little different? Does this feel right for the audience and the offer? You’re evaluating the whole campaign at once, before you’ve committed hours to final production. It’s the equivalent of a creative director pinning boards up on the wall and walking the client through the vision before anyone starts designing deliverables.

Sixty-two percent of email teams need two or more weeks to build one email campaign, according to Semrush’s 2025 research. The traditional process for getting from brief to production is slow because it’s sequential: write the brief, wait for the agency to respond, review the concept, give notes, wait for revisions, approve, then start production. With AI in the loop, the brief-to-mood-board process can happen in an afternoon. Not because you’re cutting corners, but because the feedback cycle collapses. You write the brief, the model responds instantly, you iterate in real time, and by the end of a focused session you have a campaign plan that would have taken a traditional team two weeks of back-and-forth.

Why This Step Is Worth Protecting

It’s tempting to skip the planning phase and go straight to production. When you have the ingredients and the tools, making things feels productive. Planning feels like overhead.

But every campaign I’ve produced where I skipped this step, I ended up going back and doing it anyway, just later and more painfully. You start making ads without an asset plan and you realize halfway through that your Instagram creative and your email creative are saying completely different things to the same audience. Or you build out a full set of LinkedIn ads and then realize you never thought about what happens when someone clicks through to the landing page. The campaign feels fragmented because it was never planned as a whole.

The step is also incredibly fast and cheap with AI. The brief takes maybe thirty minutes if you’ve done your upstream work. The ChatGPT asset planning conversation takes another thirty. The shot list might be another hour. The Gamma mood board, thirty minutes. You’re looking at half a day to plan a campaign that would have taken weeks at an agency, and the output is a clear, visual, stakeholder-ready document that everyone can react to before you spend any time on final production.

What to Do?

If you’re building campaigns with AI and going straight from strategy to production, try inserting this four-step planning layer.

1. Start with a campaign brief. If you don’t have a template, the GACCS format (Goals, Audience, Creative, Channels, Success metrics) is the most complete I’ve found. Comment BRIEF and I’ll send you our tried and true campaign brief skill.md file to easily create awesome briefs in Claude Cowork or the IDE of your choice. Fill it out completely. Every field. The brief should answer every question a new team member would ask before they could start working on the campaign.
2. Take the brief plus your strategic context files (positioning, ICP profiles, copy bank) and ask the LLM to build you an omni-channel campaign asset plan. Don’t just ask “what should I make?” Ask specifically: what role does each channel play? How many variants per channel? What messages map to which placements? What’s the customer journey across touchpoints?
3. From the asset plan, build a shot list. Describe every visual asset in enough detail that someone (or something) could produce it without asking follow-up questions. Reference your design tokens and brand system. Be specific about dimensions, text placement, color treatments, and mood.
4. Create a mood board. Gamma works well for this because it’s fast and visual. The goal is not perfection. The goal is to see the whole campaign at once and ask yourself: does this feel right? Does it feel like one campaign or six disconnected ads? Would I be proud to show this to a client?

The planning step is where “taste” gets applied to the system. Everything before it is preparation. Everything after it is execution. This is the moment where you decide what the campaign is actually made up of.

Subscribers can access our free Campaign Generator skill file here.

This is part of an ongoing series on go-to-market asset creation with AI. The pillar article covers the full five-stage process. Previous pieces in the series dive deep into strategic foundations, visual brand systems, and the copy bank. Next up: the production workflows for bulk ads and landing pages.