Marketing & Advertising / Google Ads

Organize keywords by search intent into themed ad groups — with specific keywords, match type recommendations, and negative keywords for each theme.
Difficulty: Intermediate → Advanced
Model: GPT-4 / Claude / Gemini
Use Case: Keyword Research, Ad Group Structuring, Intent Mapping
Updated: May 2026
Why This Prompt Exists
Most keyword research fails because it groups by topic, not by intent.

You get:

  • “best running shoes” in the same ad group as “cheap running shoes”
  • informational queries mixed with transactional (one size fits no one)
  • no match type strategy — broad match bleeding budget
  • no negative keywords — irrelevant searches triggering your ads
  • keyword lists that look impressive but don’t convert

But keyword strategy is not volume accumulation.

It is intent segmentation.

  • Search intent determines what the user needs (info vs. buy)
  • Different intents need different ads and landing pages
  • Match type saves budget from irrelevant clicks
  • Negative keywords are as important as positive ones

Without intent-based clustering, you confuse the searcher and waste budget.

This framework forces AI to be a keyword strategist who segments by intent.

The Prompt
Assume the role of a Google Ads keyword strategist who organizes by search intent, not just volume.

Your task is to generate intent-based keyword themes.

Generate 5 keyword themes, each with:

1. THEME NAME (e.g., "Problem-Solving Queries")

2. 5-7 SPECIFIC KEYWORDS
   - Use phrase match or exact match format
   - Example: "buy ergonomic chair" (exact) or "best office chair for back pain" (phrase)

3. INTENT LABEL
   - INFORMATIONAL (researching, not buying yet)
   - COMMERCIAL (comparing options, close to purchase)
   - TRANSACTIONAL (ready to buy now)
   - NAVIGATIONAL (looking for a specific brand)

4. MATCH TYPE RECOMMENDATION
   - Broad / Phrase / Exact (with rationale)

5. NEGATIVE KEYWORD
   - One keyword to exclude for this theme

INPUTS:

Product or Service:
[DESCRIBE]

Three Customer Types or Use Cases:
[E.G., "Budget-conscious first-timer" / "Quality-focused repeat buyer" / "Business purchasing for a team"]

Current Problem You're Solving (optional):
[E.G., "We're getting clicks from people looking for free templates, but we sell paid software"]

RULES:
- Do not mix intents in the same theme
- If the user's product is high-ticket (over $500), prioritize commercial and transactional intents
- Negative keywords must be specific (e.g., "free" not just "free stuff")
- Use phrase match ["keyword"] for most commercial keywords
- Use exact match [keyword] for high-intent transactional keywords
How To Use It
  • Build separate ad groups for each keyword theme — never mix intents.
  • Informational keywords need educational landing pages, not product pages.
  • Transactional keywords need product pages with clear CTAs and pricing.
  • Add negative keywords at the campaign level to prevent cross-contamination.
  • Review search term reports weekly to find new negatives.
Example Input

Product or Service: Project management software for small teams ($29/month)

Three Customer Types or Use Cases: “Freelancer juggling multiple clients” / “Small agency team of 5-10” / “Startup founder who hates spreadsheets”

Current Problem You’re Solving: “We’re getting clicks from people looking for free alternatives to Asana, but they never convert.”

Why It Works
Most keyword strategies fail because they ignore intent.

This framework improves outcomes by forcing:

  • intent-based theme naming
  • specific keyword examples (not generic categories)
  • match type recommendations (budget protection)
  • negative keywords per theme (relevance)
  • customer type alignment (audience specificity)

Great keyword strategy doesn’t find more searches — it finds the right searches.

Build Better AI Systems

Subscribe for advanced prompt engineering, AI marketing tools, Google Ads frameworks, and practical strategies for advertisers and business owners.

Leave a comment

Your email address will not be published. Required fields are marked *