Structured Campaign Creation from a Brief
Turn a standard brief into a correctly-structured campaign every single time. Built for agencies running dozens of accounts, where junior analysts do the launching and consistency is the whole point.
Works with Claude, ChatGPT, or any MCP-enabled AI tool
The problem isn't creating campaigns. It's consistency.
Every agency has a documented process for launching campaigns. And every agency knows the same truth: the human factor always shows up. A junior analyst skips a UTM, names an ad set their own way, splits the budget differently, or forgets an audience exclusion. Multiply that by 20 clients and 60+ ad accounts, and "small deviations" become a quality-control problem you can't review your way out of.
The fix is not more training or a longer checklist. It’s changing what humans produce: instead of building campaigns by hand in Ads Manager, your team fills in a standard brief — and the AI builds the campaign from it, the same way, every time. The brief becomes the single input; structure, naming, and settings are reproduced mechanically.
You standardize once. After that, a junior analyst and your most senior media buyer launch structurally identical campaigns.
Creatives → IDs → Brief → Campaign
Three steps, repeated identically for every launch, on every account.
Step 1: Upload creatives and collect the IDs
- Go to pipeboard.co/creatives, pick the client’s ad account, and drop in the images and videos.
- Copy the IDs from the "Recent uploads" panel — each creative gets an ID you’ll reference in the brief.
- Name files descriptively before uploading (e.g. clientX-promo-feed-01.jpg, clientX-promo-story-01.mp4) so the mapping to placements stays obvious.
Step 2: Fill in the standard brief
- Use the template below. This is the standardization step: objective, structure (CBO/ABO), targeting per ad set, placements, creative IDs, copy, naming, and rules.
- The brief is the contract. If it’s in the brief, it ends up in the campaign; if it’s ambiguous in the brief, fix the brief — not the analyst.
Step 3: Paste the brief into your AI chat and review
- Paste the launch prompt + filled-in brief into Claude (or any MCP-enabled tool) connected to the Pipeboard Meta Ads MCP.
- The AI creates the campaign, ad sets, and ads exactly per the brief — everything paused.
- Review the structure it echoes back, then say "publish". Review is a 2-minute diff against the brief, not a rebuild.
The standard campaign brief
Copy this into a doc or form your team fills in for every launch. Translate it to your team's language — the structure is what matters.
CAMPAIGN BRIEF (Pipeboard) - Ad account / client: act_XXXXXXXXX (client name) - Objective: Sales / Leads / Traffic / Awareness - Structure: 1 campaign, CBO or ABO · number of ad sets · total daily budget - Targeting per ad set: geo, age, gender, interests / lookalikes, exclusions - Placements: Advantage+ or manual (which creatives run in feed vs stories) - Creatives: IDs from pipeboard.co/creatives + which creative goes in each ad set - Copy per ad: headline, description, primary text, CTA, destination URL + UTMs - Naming convention: [Client]_[Objective]_[Month]_[Audience] (ad sets and ads follow the same pattern) - Rules: start paused? minimum budget per ad set? audiences must not overlap?
The launch prompt
Same prompt every time. The only thing that changes is the brief pasted under it.
Create this campaign in Pipeboard with exactly this structure and naming, creating the ad sets and ads as specified in the brief below. Start everything PAUSED. Show me the full structure (campaign, ad sets, ads, budgets, targeting, creative mapping) for confirmation before publishing. [paste the filled-in brief here]
The decisions your brief locks in
These are the choices that drift when campaigns are built by hand. Decide them once per agency (or per client), write them into the brief, and stop re-deciding.
Naming conventions
Pick one pattern and enforce it via the brief, e.g. [Client]_[Objective]_[Month]_[Audience]. Ad sets append the audience, ads append the creative variant.
Consistent names are what make cross-account reporting, bulk operations, and audits possible later — the AI reproduces the pattern exactly, so reports never break on a typo.
CBO vs ABO architecture
CBO (budget at the campaign level) lets Meta shift spend to winning ad sets — the usual default for scaling. ABO (budget per ad set) guarantees each audience gets spend — better for controlled tests.
Your brief states which one and why, plus minimum budgets per ad set. The analyst never chooses on the fly.
Feed vs Stories placements
If you run Advantage+ placements, upload both aspect ratios (1:1/4:5 for feed, 9:16 for stories/reels) and let the AI map each creative to the right slot in one ad.
For manual placements, the brief says which creative IDs are feed and which are stories — descriptive file names (…-feed-01, …-story-01) make this unambiguous.
Guardrails
Encode your safety rules directly in the brief: always start paused, never exceed the stated total budget, exclude overlapping audiences, always append UTMs.
Combined with the confirm-before-publish step, this is what makes the workflow safe to hand to a junior analyst on day one.
Pro tips
- Use one chat per campaign launch and keep it short — long conversations make the AI more likely to drift from the brief.
- Never publish directly: the "confirm the structure before publishing" line in the prompt is your review gate. The reviewer compares the echoed structure against the brief, not against memory.
- When a launch reveals a gap (a missing rule, an ambiguous field), fix the brief template — that's how the process improves for every future launch.
Related Guides
Continue learning with these recommended guides.