SEM Search Intent Classifier
Stage-rank every keyword in the buyer journey and get a clear SEM value score to focus your budget
What This Prompt Does
Traditional intent classification tells you the category. This version goes further — it places every keyword in a buyer stage (early research, comparison, ready to buy, looking for a provider now) and assigns a strong/medium/weak SEM value, so you know exactly which terms deserve budget.
When to Use It
- •Triaging a keyword list before building campaigns
- •Deciding which existing keywords should stay and which to pause
- •Explaining to a client why some keywords are not worth bidding on
- •Prioritizing limited budget toward the highest-intent phrases
The Prompt Template
Act as a paid search analyst. Your goal is to classify these keywords by buying intent so ad spend can be focused on the terms most likely to convert. Context: - Business type: [BUSINESS TYPE] - Keyword list: [PASTE KEYWORDS] - Conversion goal: [LEADS, SALES, CALLS] Task: For each keyword, identify: 1. the likely search intent 2. the likely buyer stage 3. whether the keyword is a strong paid search target Use these buyer stages: - early research - comparison - ready to buy - looking for a provider now Process: 1. Read each keyword. 2. Determine what the searcher is trying to do. 3. Place the keyword in the buyer journey. 4. Mark whether it is a strong, medium, or weak SEM keyword. Constraints: - Use simple language. - Keep each explanation to 1 sentence. - Do not overcomplicate the categories. - Focus on business value, not just relevance. - Favor recommendations that reduce wasted spend and improve conversion quality, not just click volume. Output format: Return a table with these columns: - Keyword - Intent - Buyer Stage - SEM Value - Reason
How the Prompt Is Structured
Four Buyer Stages, Not Five
Limiting the taxonomy to four stages forces the model to commit. More categories sound thorough but produce fuzzy decisions. "Looking for a provider now" is deliberately separated from "ready to buy" because the two trigger different campaigns.
Strong / Medium / Weak — Not a Score
Three buckets is enough. Numeric scales (1–10) encourage false precision; three labels force a real decision and make spreadsheet filtering easier.
"Business Value, Not Just Relevance"
A keyword can be perfectly relevant and still a waste of budget. This constraint pushes the model away from the relevance trap that burns paid budgets on informational content.
Single-Sentence Reasons
A one-sentence reason is enough to defend the classification to a client or teammate. Longer explanations balloon the table and obscure the actual verdict.
Example Output
| Keyword | Intent | Buyer Stage | SEM Value | Reason |
|---|---|---|---|---|
| best CRM for small business | commercial | comparison | medium | Comparing options but may not be ready to buy. |
| emergency plumber near me | local / transactional | looking for a provider now | strong | Urgent local intent, ready to call. |
| how does CRM software work | informational | early research | weak | Educational query unlikely to convert on a paid click. |
| buy hubspot | navigational | ready to buy | weak | Wrong brand — competitor traffic, avoid unless doing conquest. |
Tips for Better Results
Work in Batches of 50
Long lists dilute judgment. Classify 50 keywords at a time, spot-check, and then run the next batch.
Filter to "Strong" First
Your first campaign should contain only strong-value keywords. Medium and weak go into separate groups (or the bench) for later.
Push Back on "Medium"
If a keyword lands in medium, ask the AI what would push it to strong. That follow-up reveals useful modifiers you can test.
Feed Into Campaign Structure Next
Pass the labeled table to the Campaign Structure prompt to turn buyer stages into ad group themes.
Spend Where Buyers Actually Search
We help you prune paid keyword lists down to the phrases that actually produce revenue.