Inside this resource.
9 sectionsA map of the published document. Choose a section to start reading.
- 01Put this guide to work
- 02Before You Start: The Context Difference
- 03Prompt 1: Search Term Waste Finder
- 04Prompt 2: Competitor Ad Copy Analyzer
- 05Prompt 3: Ad Copy Generator With Built-In Quality Filter
View 4 more sections
The prompt is one step in the review.
Each prompt needs permitted inputs and an operator checking the output. A generated recommendation is not an account change or a measured saving.
- Export
Select the actual source, columns and date window.
- Add context
Supply product facts, goals and protected brand terms.
- Inspect
Recalculate the draft and check each proposed action.
- Decide
The account owner chooses the permitted next step.
Retain the inputs and review notes so the next draft can be checked against the same evidence.
Read the full source: Before You Start: The Context Difference · Prompt 1: Search Term Waste Finder · Prompt 2: Competitor Ad Copy Analyzer
Jump to a section
Put this guide to work
Try search-term review, competitor copy analysis and ad drafting with appropriate context and human review.
Your task: Choose a starter prompt. Record the source, date range and reviewer beside the result. Leave anything you have not checked marked Not checked.
Related work: Map buyer searches before you build campaigns. · 3 Essential Scripts Starter Pack
Before You Start: The Context Difference
Most people paste "write me 15 Google Ads headlines for [product]" and get 15 vague, interchangeable headlines back. That's not an AI workflow. That's asking a question.
These prompts work because they include context. Each one primes the AI with specific data about your account, your competitors, and your customers before asking it to produce anything.
Context helps make output relevant, but model behavior and verification also matter. Share only account data you are permitted to process in the chosen tool.
Prompt 1: Search Term Waste Finder
What it replaces: 2-3 hours of manual search term report analysis What it produces: Query-review candidates with observed cost, grouped by theme
Setup (5 minutes)
- Open Google Ads
- Navigate to Keywords > Search Terms
- Set date range to last 30 days
- Add columns: Search term, Impressions, Clicks, Cost, Conversions, Conversion Value
- Export as CSV
The Prompt
Copy this prompt and paste your CSV data where indicated:
You are a Google Ads specialist analyzing search term data. I'm going to give you a search term report. Your job is to flag queries for human review and propose negative-keyword candidates. Thresholds alone do not prove waste. Respect conversion lag, business relevance and the supplied date range; missing data is unknown.
MY ACCOUNT CONTEXT:
- Business type: [YOUR BUSINESS TYPE - e.g., "e-commerce selling organic skincare"]
- Target CPA: $[YOUR TARGET CPA]
- Products/services: [LIST YOUR MAIN PRODUCTS]
- Branded terms to protect: [YOUR BRAND NAME AND VARIATIONS]
INSTRUCTIONS:
1. Identify every search term that has:
a. Zero conversions AND spent more than $[YOUR TARGET CPA x 2]
b. CTR below 1.5% with 100+ impressions (poor relevance signal)
c. Informational intent ("how to," "what is," "why") in commercial campaigns
2. Group the review candidates by theme (e.g., "competitor brand terms," "informational queries," "irrelevant category," "geographic mismatch")
3. For each group, recommend:
- Specific negative keywords to add (phrase match or exact match)
- Whether to add at campaign level or account level
- Observed cost in the supplied period, distinct from any explicitly hypothetical forecast
4. Flag any search terms that are converting well but might be undervalued (low CPA, low spend - these are expansion opportunities)
5. Sum observed cost across unique flagged terms without double counting. Label it review-candidate cost, not proven waste or savings.
FORMAT: Output as a table with columns: Search Term | Cost | Conversions | Issue | Recommended Negative | Match Type | Level (Campaign/Account)
SEARCH TERM DATA:
[PASTE YOUR CSV DATA HERE]
What to Expect
Expect a draft list of review candidates and proposed exclusions. Recalculate costs and validate relevance; the prompt establishes no average savings.
Critical step after running: Don't blindly add every recommendation. Review each negative keyword to make sure it won't block legitimate queries. AI occasionally flags terms that look irrelevant but actually convert at low volume.
The Manual Version Takes 2-3 Hours Because:
- Sorting 500+ search terms by multiple criteria
- Mentally classifying each term's intent
- Grouping related waste terms into themes
- Calculating observed candidate cost per theme
- Deciding phrase match vs. exact match for each negative
Time the complete workflow, including export, validation and correction. Review effort varies with report size and quality.
Prompt 2: Competitor Ad Copy Analyzer
What it replaces: 1-2 hours of manual competitor ad research What it produces: Messaging angle breakdown, positioning gaps, counter-headline ideas
Setup (10 minutes)
- Go to Google Ads Transparency Center (adstransparency.google.com)
- Search for 3-5 competitor brand names
- For each competitor, copy the text of 5-10 of their active Search ads (focus on ads that have been running longest)
- Note approximate ad age if visible (months running)
The Prompt
You are a competitive intelligence analyst for Google Ads. I'm going to give you ad copy from my top competitors. Analyze their messaging strategy and identify positioning gaps.
MY BRAND CONTEXT:
- My brand: [YOUR BRAND NAME]
- My primary value proposition: [WHAT MAKES YOU DIFFERENT]
- My category: [YOUR PRODUCT CATEGORY]
- My strongest proof point: [YOUR BEST METRIC - e.g., "4.8 stars from 2,300 reviews" or "12 years in business"]
COMPETITOR ADS:
[For each competitor, paste their ads in this format:]
COMPETITOR: [Name]
Ad 1: [Headline] | [Description] | Running approximately [X months]
Ad 2: [Headline] | [Description] | Running approximately [X months]
[Continue for 5-10 ads per competitor]
ANALYSIS INSTRUCTIONS:
1. For each competitor, classify their ads across these 10 messaging angles:
- Price/Value
- Quality/Premium
- Urgency/Scarcity
- Social Proof
- Authority/Expertise
- Convenience/Ease
- Guarantee/Trust
- Free Offer
- Comparison
- Newness/Innovation
Show the percentage of their ads using each angle.
2. Build a messaging matrix showing all competitors side by side.
3. Identify:
- RED OCEAN angles (above 40% average usage - oversaturated)
- CONTESTED angles (15-40% usage)
- BLUE OCEAN angles (below 15% usage - positioning opportunities)
4. For the top Blue Ocean opportunity, generate 5 counter-positioning headlines that:
- Are under 30 characters each
- Would NOT sound like something any of these competitors would say
- Include at least one specific, verifiable claim
- Avoid these banned words: Premium, Innovative, Experience, Discover, Unlock, Elevate, Transform
5. For each counter-headline, explain the "shadow weakness" of the competitor angle it counters.
FORMAT: Start with the messaging matrix, then gaps analysis, then counter-headlines with reasoning.
What to Expect
You'll get a clear picture of what messaging territory is crowded, what's open, and specific headline ideas that occupy the open territory.
The most valuable output is usually the Blue Ocean identification. Most brands compete on the same 2-3 angles as everyone else without realizing there are entire positioning territories nobody has claimed.
How to Use the Output
- Validate the angle classifications by spot-checking a few ads yourself
- Use the Blue Ocean angles to inform your next ad refresh
- Test 2-3 of the counter-headlines in your next RSA
- Re-run this analysis quarterly to track shifts
Prompt 3: Ad Copy Generator With Built-In Quality Filter
What it replaces: 1-2 hours of writing + editing ad copy What it produces: 25 scored headline candidates for one RSA, pre-filtered for AI tells
Setup (5 minutes)
- Identify the keyword cluster this RSA will target (3-5 keywords with shared intent)
- Gather: your top-performing existing headlines (with CTR/conversion data if available), 2-3 customer review quotes about the problem your product solves, your primary offer/CTA
The Prompt
You are a Google Ads copywriter who has been trained to avoid AI-sounding copy. Generate headline candidates for a Responsive Search Ad, then self-score each one.
KEYWORD CLUSTER:
[LIST YOUR 3-5 TARGET KEYWORDS]
BRAND CONTEXT:
- Brand: [YOUR BRAND NAME]
- Product: [SPECIFIC PRODUCT OR SERVICE]
- Primary differentiator: [WHAT MAKES YOU DIFFERENT]
- Offer: [CURRENT OFFER - e.g., "Free shipping over $49" or "30-day trial"]
- Strongest proof: [YOUR BEST SPECIFIC METRIC]
CUSTOMER LANGUAGE (from real reviews/forums):
- "[PASTE A CUSTOMER QUOTE ABOUT THEIR PROBLEM]"
- "[PASTE ANOTHER CUSTOMER QUOTE]"
- "[PASTE A THIRD CUSTOMER QUOTE]"
TOP-PERFORMING EXISTING HEADLINES (if available):
- "[YOUR BEST HEADLINE]" - CTR: X%
- "[YOUR SECOND BEST]" - CTR: X%
GENERATION RULES:
1. Generate 25 headline candidates. Each must be under 30 characters.
2. Mix three emotional approaches:
- First-Person Pain (articulate the reader's problem): 8 headlines
- Specific Moment (capture a situation they recognize): 8 headlines
- Emotional Contradiction (tension between two truths): 9 headlines
3. HARD RULES - auto-fail any headline that:
- Uses em dashes
- Contains these words: Experience, Discover, Unlock, Elevate, Transform, Unleash, Premium (without evidence), Innovative, World-class
- Uses "Your journey to..." or "Take X to the next level" or "Everything you need"
- Could be said by any competitor with zero modification
4. For each headline, self-score on a 100-point rubric:
- Visualizable (can you picture it?): 0 or 10
- Falsifiable (could it be proven wrong?): 0 or 10
- Unique (would a competitor NOT say this?): 0 or 10
- Relevance to keyword cluster: 0-20
- Clarity in under 30 chars: 0-15
- Proofability (claims backed by evidence): 0-15
- Emotional punch: 0-10
- Customer language used: 0-10
5. Sort by score descending. Mark the top 15 as SELECTED and the bottom 10 as CUT.
FORMAT: Table with columns: # | Headline | Chars | Approach | Score | Status (SELECTED/CUT) | Reasoning
What to Expect
You'll get 25 headlines with scores. The top 15 become your RSA candidates. The scoring makes it easy to see why some work and others don't.
After running this prompt:
- Verify every specific claim (review counts, ratings, stats). Round review counts DOWN, not up.
- Check that no two of your selected headlines start with the same word or follow the same structure.
- Count characters yourself. AI sometimes miscounts.
- Read the top 15 out loud. Do any sound robotic? If yes, rephrase in your own words while keeping the angle.
What These Three Prompts Demonstrate
Each prompt follows the same structure that makes AI useful for ads:
Context first. Every prompt starts with brand-specific information. This is the single biggest factor in output quality.
Structured instructions. Not "write headlines" but "generate 25 candidates following these specific rules, then score each one against this rubric." Structure prevents generic output.
Built-in quality gates. The banned word list, the scoring rubric, the character limits. These constraints force AI to work harder and produce better results.
Human validation required. Every prompt includes a "what to do after" section. AI generates. You validate. The combination is what works.
Measure the Time on Your Own Workflow
| Task | Manual Time | With These Prompts |
|---|---|---|
| Search term waste analysis | 2-3 hours | 15 minutes (including review) |
| Competitor ad analysis | 1-2 hours | 20 minutes (including setup) |
| RSA headline generation | 1-2 hours | 15 minutes (including validation) |
| Total | 4-7 hours | 50 minutes |
The table is an illustrative timing scenario, not observed savings. Record your own baseline and include setup, validation, correction and maintenance before estimating net time saved.
Going Further
These three prompts are the starting point. The complete AI workflow system includes:
The Full Prompt Library:
- 7-Prompt Keyword Research Sequence (organizes research, validation and campaign mapping)
- Feed Optimization prompts (product title rewriting at scale)
- Landing page copy generation (section-by-section)
- Performance narrative generation (turns raw data into action plans)
- Campaign-overlap investigation
- Anomaly detection for catching tracking breaks
The Context Stack:
- Brand brief template (200-400 words)
- Competitor map framework (5-10 competitors)
- Customer language file builder
- Performance history log for feeding results back into the system
The Anti-AI Copy Filter:
- Complete banned word and phrase list
- 5-point emotional validation checklist
- Before/after transformation examples
- Anti-template enforcement rules
The 8-Minute Weekly Audit:
- 6-step workflow for a bounded review and a queue of deeper investigations
- Automated anomaly detection setup
- Campaign-overlap review methodology
- Performance narrative generation
Want the Complete Agentic System?
These 3 prompts demonstrate the principle. The Google Ads AI Agentic System (see the current product page for price and license terms) gives you the full toolkit:
- The complete context stack with templates and examples
- 7-prompt keyword research sequence with source checks and operator review
- Overgenerate-and-score ad copy method with 100-point rubric
- AI-powered feed optimization with keyword-coverage review
- Landing page copy generation with awareness-level mapping
- Creative production system with 7-section prompt structure
- The complete anti-AI copy filter (structural tells, semantic tells, banned words)
- 8-minute weekly audit workflow
- Workflow automation setup (n8n/Make integration)
- Implementation roadmap for planning setup and validation work
The quick start gets you moving. The full system builds the competitive advantage.
Get the Google Ads AI Agentic System at tegra.co/store/google-ads-ai-agentic-system
Choose who handles the next step.
Use the reference yourself, book a focused paid working session, or have Tegra handle the ongoing channel work.
Search, Shopping, Performance Max and Demand Gen campaigns, plus the research, product feeds, static and video creative, and landing, offer and pre-sale pages behind them.
Explore managed Google Ads → · See paid working sessions →