Home / Blog / Create Google Ads in 4 Prompts (Premium)
Premium Features

Create a Google Ads Campaign in 4 Claude Prompts (Premium Walkthrough)

May 28, 2026 8 min read

A new Google Ads campaign takes a competent PPC marketer 25-45 minutes to launch properly: campaign settings, ad groups, keywords, ad copy, extensions, conversion tracking links. With Claude + 1ClickReport Premium ($99/mo), the same launch takes 4 prompts and roughly 8 minutes — and you can do it from the same Claude window where you analyse performance. Here's the exact 4-prompt walkthrough with a real example.

AR
Written by Allan Rufus
Performance Marketing Lead · LinkedIn

Performance marketing lead with 8+ years running paid media for DTC, SaaS, and high-ticket coaching brands.

What you need to start

  • 1ClickReport Premium plan ($99/mo) — Pro is read-only; Premium has the campaign-creation tools
  • OAuth-connected Google Ads account with manage permissions (not view-only)
  • A target keyword or keyword cluster, a daily budget, and a landing page URL
  • Claude Desktop, Claude Code, or another MCP-compatible client

Total prep time once you have a campaign idea: about 2 minutes.

Prompt 1: Plan the campaign

I want to launch a new Google Ads Search campaign targeting the
keyword 'marketing analytics dashboard' for a B2B SaaS audience.
Budget: $50/day. Target country: US. Landing page:
https://www.1clickreport.com/dashboard-builder

Recommend:
1. Campaign type and bid strategy for this budget level
2. Ad group structure (single ad group or multiple?)
3. Keyword list with match types
4. Negative keywords to start with
5. Whether I should use Sitelinks, Callouts, or Structured Snippets

What you get back: a full campaign plan. For the example above, Claude typically recommends:

  • Search campaign, Maximize Conversions bid strategy (no conversion history → Manual CPC alternative offered)
  • 2 ad groups: "marketing analytics dashboard" (exact + phrase) and "marketing dashboard alternatives" (related intent)
  • ~15 keywords across both ad groups with mixed match types
  • ~20 negative keywords to start (industry-specific, geography, irrelevant intent)
  • Yes to all three extensions, with suggested copy

This is the plan. You review, adjust if needed, then commit.

Prompt 2: Draft the ad copy

For the campaign plan above, draft 3 responsive search ad variations
for each ad group. For each ad:
- 15 headlines (max 30 chars)
- 4 descriptions (max 90 chars)
- Pinned headlines: pin "1ClickReport" to position 2

Voice: direct, no superlatives, lead with concrete benefit. Include
the keyword in 50% of headlines.

What you get back: 6 complete ad drafts (3 per ad group × 2 ad groups). Each with 15 headlines + 4 descriptions, properly within character limits, with pinning applied.

Sample headlines from a real run:

  • "Marketing Analytics Dashboard"
  • "60-Second Dashboard Setup"
  • "GA4 + Meta + Google Ads"
  • "Connect Data in 1 Click"
  • "$25/mo. Free Trial."

Review, swap any you don't like, add brand-specific ones if needed. Approve before commit.

Prompt 3: Build the campaign in Google Ads

Build the campaign in my Google Ads account based on the plan + ads
above. Set up:
- Campaign in PAUSED status initially (I want to review before launch)
- Both ad groups with the keywords and match types
- All 6 responsive search ads
- The 20 negative keywords applied at campaign level
- Sitelinks: Pricing, Free Trial, Customer Stories, Docs
- Callouts: "60-second setup", "Connect 5+ data sources", "No credit card"

Confirm what's been built with a structure summary.

What happens: Claude calls create_google_campaign, create_google_ad_group, manage_google_keywords, create_google_ad, set_campaign_assets, and related tools in sequence. Output is a structural summary showing exactly what was created.

The campaign appears in Google Ads UI within seconds. Because it's in PAUSED status, nothing runs yet.

Prompt 4: Final review before activating

Pull the campaign you just built and audit it against best practices:

1. Are all ads at "Good" or "Excellent" ad strength?
2. Is conversion tracking properly linked?
3. Are bid limits or target CPA settings reasonable for the $50/day
   budget?
4. Is location targeting set correctly?
5. Any settings I'd typically miss?

Recommend whether to activate now or fix issues first.

What you get back: a self-audit. Common findings:

  • "Ad 1 in Ad Group 2 is at 'Average' strength — add 3 more unique headlines"
  • "Conversion tracking linked correctly to your /auth.html signup event"
  • "Recommend setting bid limit at $4 to prevent runaway CPCs in first 7 days"
  • "Location targeting: US only. Recommend excluding Hawaii + Alaska if not relevant"

You fix the flagged items in the conversation ("update ad 1 to add these headlines: ..."), Claude executes, then you activate the campaign.

Total time from "I want to launch a campaign" to "campaign is live": ~8 minutes.

Where the time savings come from

ActivityManual timeClaude time
Campaign strategy planning5-10 min30 sec
Writing 15 headlines + 4 descriptions × 6 ads15-25 min1 min
Building campaign UI (ad groups, keywords, settings)10-15 min1 min
Adding extensions, negative keywords, structured data5-10 minincluded above
Final review3-5 min1 min
Total38-65 min~8 min

The largest single saver is ad copy. Writing 6 RSAs by hand is the single most painful part of campaign launch for most marketers. Claude does it in seconds, you edit lightly.

What you should still do manually

  • Conversion event configuration. Set this up once in Google Ads UI per account. Claude can verify it's linked but shouldn't configure it from scratch.
  • Strategic decisions about which campaigns to launch. Claude can suggest plans; you decide whether to invest.
  • Account-level settings. Auto-tagging, time zone, billing — these should be set in the UI once, not per campaign.
  • The first real budget changes after launch. Let the campaign run for 5-7 days before making bid or budget changes — early signal is noisy, and Claude will optimize on noise just as readily as a human.

Is Premium worth $99/mo just for campaign creation?

Cost-justification math: if you launch 1+ new campaign per week at 40 minutes each saved, that's 40+ minutes/week. At a $50/hr opportunity cost, that's $33/mo in time saved — close to Pro's full price.

If you launch 5+ campaigns/month, you're saving 3-5 hours/month, easily covering the $99 jump from Pro.

For agencies launching 10+ campaigns/month across clients, the math becomes obvious — Premium pays back in the first day of every month.

For solo founders running 1-2 always-on campaigns with rare new launches, Pro is sufficient. Use Pro to analyze, switch to Premium temporarily when you're in a campaign-building sprint.

Frequently Asked Questions

Can Claude actually create a campaign that goes live immediately?

Yes, but the recommended pattern is to create it in PAUSED status, review the full structure, then manually activate. This catches Claude's occasional misinterpretations before money starts flowing.

What if Claude builds the campaign wrong?

Same recourse as any tool error — pause the campaign, audit it, fix the specific issues (or delete and recreate). The Google Ads API actions are reversible; nothing Claude does that you can't undo.

Does this work for Performance Max campaigns?

Partially as of mid-2026. PMax has limited API exposure compared to Search — campaign creation works, but some asset group customization still requires the UI. For pure Search campaigns, full coverage.

What about Shopping campaigns?

Search and Shopping both supported via the campaign creation tools. Shopping requires a Merchant Center feed connection that you'd set up separately in Google Ads first.

Can I create campaigns for multiple clients in one Claude session?

Yes — switch active client context between prompts and create campaigns for each. The multi-account setup is the same as analytics — see the multi-client MCP setup walkthrough for the broader pattern.

What's the risk of Claude making a costly mistake?

Real but manageable. Premium has built-in confirmations for destructive actions and recommends paused-then-activate workflows. The biggest risk isn't Claude executing wrong — it's you approving Claude's recommendations without reading them carefully. Treat every Claude-suggested campaign launch the way you'd treat a junior team member's first launch: review the structure before activating.