Skip to main content

Paid media · build guide

Campaign Launch Kit

Brief in, ready-to-launch campaign kit out — and nothing spends until you approve it.

Path A · Manual extracts Path B · MCP integration Guardrail-first

Campaign Launch Kit assembles a complete, ready-to-upload campaign from a one-page brief: a strategic brief, an ad group structure organized by theme, intent tier, and match type, three copy variants per ad group, landing page outlines mapped to each group's intent, a documented UTM taxonomy, and a pre-launch QA checklist covering tracking, copy, targeting, budget, and landing page parity. Everything arrives paused — no spend starts until a human enables it.

Every launch feeds the same templates and a persistent instruction file that records your offer, audience, naming conventions, and guardrails — so kit two is faster and more consistent than kit one. Setup drops from days to hours, and clean UTMs from day one mean every downstream report can be trusted.

What it moves: Campaign setup time drops from days to hours, with every launch shipping a consistent structure, clean UTMs, and a passed QA checklist.
Path A · Manual

You run the kit from templates: a brief document, a Google Sheets workbook, and Claude in a chat window. There are no system connections — you paste the brief in, paste the kit out, and upload through Google Ads Editor yourself. It suits teams launching a campaign or two a month who want the speed without granting any API access.

Path B · Integrated

Your project management tool, ad account, and asset storage connect to an agent that runs the whole kit on a trigger: a launch-request task goes in, a paused and QA'd campaign comes out. The agent acts within hard caps and logs every step; you approve before anything spends. It suits teams launching often enough that copy-paste is the bottleneck.

Prerequisites.

A documented launch process — the steps and sign-offs your team already follows; the automation encodes it, it can't invent it
Consistent naming conventions for campaigns, ad groups, and UTMs
Offer, audience/ICP, and brand tone-of-voice documentation the kit generator can draw from
A project management tool where launch requests live (needed for the MCP path)
Access to your ad account — Google Ads Editor for the manual path, API access for the MCP path
New to the stack? Set up your environment first — data access, workspace, integration platforms, and the agent layer are covered once in Environments & tooling.

Build it with manual data extracts.

You run the kit from templates: a brief document, a Google Sheets workbook, and Claude in a chat window. There are no system connections — you paste the brief in, paste the kit out, and upload through Google Ads Editor yourself. It suits teams launching a campaign or two a month who want the speed without granting any API access.

  1. Fix the one-page brief format

    Create a one-page brief document with fixed fields: offer, audience and ICP, intent clusters, competitive angle, and success metrics. Fill one out per campaign. A complete brief is the input contract — if a field is blank, the kit will be vague, so treat blanks as blockers.

  2. Set up the kit workbook

    Create a Google Sheets workbook with one tab each for ad group structure, copy variants, landing page outline, UTM builder, and QA checklist. Encode your naming conventions as data-validation rules or formulas so a wrongly named campaign shows up red before it ever reaches the account.

  3. Generate the kit with Claude

    Paste the brief and your naming conventions into Claude and ask for the ad group structure organized by theme, intent tier, and match type, plus three copy variants per ad group and a landing page outline mapped to each group's intent. Paste the output into the workbook tabs. Edit in the sheet, not in the chat — the workbook is the source of truth.

  4. Build UTMs in the sheet

    The UTM tab constructs final URLs by formula from your documented convention — source, medium, campaign, content, term. Nobody hand-types a UTM. Copy the finished URLs into the copy tab so every ad row carries its tracking.

  5. Run the QA checklist

    Work the checklist tab row by row: tracking in place, copy within character limits, targeting matches the brief, budget matches the plan, landing page matches the ad's promise. Any blocker stops the launch until it is resolved; warnings get a note for the reviewer.

  6. Upload paused via Google Ads Editor

    Import the structure and copy through Google Ads Editor with every campaign paused. A second person reviews in-account against the workbook, then enables. Because the upload is paused, backing out is as simple as deleting the draft — nothing has spent.

Cadence: Run once per campaign launch; with templates in place, a full cycle takes hours instead of the days a hand-built campaign takes.

Integrate it with MCP connections.

Your project management tool, ad account, and asset storage connect to an agent that runs the whole kit on a trigger: a launch-request task goes in, a paused and QA'd campaign comes out. The agent acts within hard caps and logs every step; you approve before anything spends. It suits teams launching often enough that copy-paste is the bottleneck.

Project management tool (Asana, Jira, or similar — API access or an MCP server) for launch requests and approvals
Google Ads API (or an MCP server for Google Ads), scoped to creating campaigns in a paused state
Google Sheets or Drive for the kit archive and audit log
GA4 (read-only) to verify tracking fires and UTM data lands correctly
  1. Connect the systems

    Give the agent your project management tool for launch tasks and approvals, the Google Ads API (or an MCP server for Google Ads) scoped to creating campaigns in a paused state, Google Sheets or Drive for the kit archive, and GA4 read-only for tracking checks. Wire the trigger with n8n, Zapier, Make, or Claude with MCP servers directly.

  2. Encode your launch process as the agent's instructions

    Write a persistent instruction file holding your offer, audience, naming conventions, budget caps, and required QA gates. This file is the governance layer: the agent refuses to build from an incomplete brief and cannot act outside what the file permits. It carries forward, so every future launch starts from the same guardrails.

  3. Agent assembles the kit

    When a launch task is tagged in your PM tool, the agent reads the brief fields and generates the strategic brief, ad group structure by theme, intent tier, and match type, three copy variants per group, landing page outlines, and UTMs built strictly from your taxonomy. Everything is written to the workbook archive before anything touches the ad account.

  4. Agent runs automated QA

    The agent validates every name against your convention, checks UTM parity between ads and landing pages, confirms budgets sit within cap, and verifies via GA4 that tracking is in place. Blockers halt the run; warnings are collected for the reviewer rather than silently cleared.

  5. Paused upload and approval gate

    The agent creates the campaigns through the ads API in a paused state with the capped daily budget already set, then posts the QA results and a summary of exactly what was built to the PM task. A named approver reviews and enables in-account. Nothing spends on the agent's authority alone.

  6. Log and improve

    Every run appends to the audit log: inputs, generated assets, QA outcomes, who approved, and when spend was enabled. After launch, feed what worked back into the instruction file so the next kit starts smarter without loosening any guardrail.

Run it safely.

Claude with MCP servers — generates the brief, structure, and copy, and runs the QA loop on the connected path
Google Sheets — the kit workbook, formula-driven UTM builder, and QA checklist on both paths
Google Ads Editor — bulk paused uploads on the manual path, fully reversible before anything runs
Google Ads API — programmatic paused campaign creation on the MCP path
n8n, Zapier, or Make — wires the PM-tool launch trigger to the agent run
GA4 — confirms tracking fires and UTM data arrives clean
  • Every campaign is created paused; only a named human approver enables spend
  • Daily budget caps are set at creation and the agent cannot exceed them — raising a cap is a human-only change
  • QA must pass with zero blockers before an approval request goes out; warnings are surfaced to the reviewer, never cleared silently
  • Every run writes an audit log — inputs, generated assets, QA results, approver, enable time — and a paused upload can be deleted with nothing spent

Want new build guides as they ship?

Optional — every guide on this site is free either way. Leave your details and we'll email you when new automations join the library, nothing else.

We only email when there's something worth building — every email has a one-click unsubscribe, and you can withdraw consent anytime. Consent requested by Campaign Automation AI, Caledon, Ontario · hello@campaignautomation.ai.

Prefer it built for you?

We stand this automation up as a governed sprint on the stack you already run — scoped, built, and handed over with the guardrails configured. No new platform, no migration.

Start with the Audit → Request proposals →