Agent 05 · Per Launch

The Campaign Builder

A launch brief is five answers: goal, money, subject, audience, guardrails. Between that brief and a live campaign sit forty mechanical decisions — and that is where launches go wrong. This agent handles the forty, documents its reasoning on every one, and stops at a launch card until a named human approves.

Get this skill — free in the packMageCloud growth services
Paul Ryazanov on stage — one of 15+ international conference appearances

Paul Ryazanov on stage — one of 15+ international conference appearances

What it takes off your team’s plate

Launch mechanics that used to eat specialist hours:

Structure decisionsMatch types, ad group themes, priority tiers, asset group splits — decided consistently instead of by whoever is launching.
Asset assemblyTwelve-plus headline angles covering price, proof, shipping, and objections — with negatives attached at creation, not after the first wasted week.
Platform rulesCBD, pharma, and alcohol constraints checked before building, not discovered at disapproval.
DocumentationA record of what was chosen and why, so the next launch starts smarter.

How a run works

Brief to launch card:

  1. Require the brief. Goal, budget and duration, products or video, audience in human terms, guardrails. Missing answers get asked, not assumed.
  2. Build by type. Search: one theme per ad group, deliberate match types. Shopping: margin tiers and brand-term cascades. PMax: every asset slot filled, brand exclusions on, URL expansion off by default. YouTube: engagement goals, custom search-behavior segments, dollar-a-day test budgets scaled to store size.
  3. Verify tracking before budget. No campaign gets money against a conversion action that has not fired recently.
  4. Present the launch card. Every setting with its reasoning in one table: name, type, budget, bidding, targeting, assets, negatives, tracking check. The human approves, edits, or rejects.
  5. Launch and document. After approval only. First checkpoint at 72 hours — delivery sanity, no performance judgments on three days of data.

✋ The approval gate

Nothing goes live on its own:

  • The launch card requires a named human approval — always
  • Tracking verified before budget is assigned
  • Brand negatives attached to every non-brand campaign at creation
  • Every build documented: what was chosen, why, and what it cost

What you get

A launch-ready campaign, held at the approval gate
The Launch Card — every setting with its reasoning
BUILD-[client]-[campaign].md — the documented build
A 72-hour delivery checkpoint, scheduled
Ecommerce Camp — the UK practitioner community Paul founded
Ecommerce Camp — the UK practitioner community Paul founded

This is the same pattern Dennis Yu’s team demonstrated publicly when an agent built and launched a complete YouTube campaign inside Google Ads for under a dollar of compute — generalized to the campaign types an e-commerce agency launches every week.

I have written about this from the trenches:

Run it yourself, or bring my team in

The skill is free in the pack — install it and start this week. If your store does seven figures and you want this running inside a full revenue-optimization engagement, my team at MageCloud handles it end to end.

Download the packBook a call
© Paul Ryazanov · paulryazanov.com · MageCloud · Ecommerce CampBuilt in partnership with Dennis Yu · how this system was built