Buyer-intent keywords, the AI-native way
The Keyword research note (previous step) hands you a long list. This note is how you choose from it: AI tags every keyword's intent in one pass, and your margin numbers decide the rest. A term with 500 buying-intent searches beats one with 50,000 browsers.
The intent ladder
Every query sits on a ladder — the higher the rung, the closer to the wallet. For a yoga-mat store:
| Rung | The searcher is… | Real query shapes | Belongs on |
|---|
| Informational | learning | how thick should a yoga mat be; how to clean a yoga mat | blog post |
| Commercial | comparing | best yoga mat for hot yoga; manduka vs lululemon mat; cork yoga mat review | comparison or best-of page |
| Transactional | buying | buy cork yoga mat; extra long yoga mat 84 inch; manduka pro sale | product or collection page |
Modifier cheat-sheet: how / what / why = informational; best / vs / review / for [use-case] = commercial; buy / price / sale / exact model or size = transactional. Let blog posts own the top of the funnel; let product and collection pages own the bottom.
Fastest path: one prompt, end to end
Paste your keyword list into ChatGPT or Claude with this:
🤖 AI prompt — paste into ChatGPT / Claude
You are an SEO strategist for an ecommerce store.
Store sells: [your products in one sentence]
Keyword list: [paste keywords, one per line, with volume and difficulty if you have them]
Do this and return ONE markdown table:
1. Label each keyword's intent: Informational, Commercial, or Transactional. Use the modifiers (how/what/why = informational; best/vs/review = commercial; buy/price/sale/exact model = transactional) plus common sense.
2. Flag every buyer term (Commercial or Transactional) with a star.
3. For each buyer term, name the product or collection in my store it would sell, and the page type it belongs on (product, collection, comparison, blog).
4. If a keyword has no volume number, leave the Volume column blank. Do not guess or estimate any number — every figure must come from a tool you opened or a source you can cite next to it. If you can browse, read real volumes from Google Keyword Planner (ads.google.com) or the Ahrefs Free Keyword Generator (ahrefs.com/keyword-generator); if you cannot browse, leave Volume blank and tell me to fill it from one of those tools.
5. Sort the table: Transactional first, then Commercial, then Informational.
Columns: Keyword | Intent | Buyer? | Volume | Sells which product | Page type
Or re-score your list in 5 steps
- Start from the sheet you built in Keyword research — or export the last 12 months of queries from Google Search Console → Performance.
- Let AI tag intent. Paste the whole list: "Label each keyword Informational, Commercial or Transactional and flag the buyer terms." Spot-check 10 by hand; fix any wrong ones.
- Add margin. For each buyer term, note which product it sells and that product's gross margin. Same volume, very different value: a term selling a 70%-margin product beats one selling a 20%-margin product.
- Score. Intent weight (Transactional 3, Commercial 2, Informational 1) × gross margin %. Volume only breaks ties.
- Re-rank and map. The top 10 each get a page: transactional → product or collection page, commercial → comparison page, informational → blog post. Done = the re-ranked sheet with a page named next to each of the top 10.
Worked example — example math, your numbers will differ
Say four yoga-mat terms come out like this (volumes are placeholders — pull real ones from Keyword Planner):
| Keyword | Intent (weight) | Margin of product it sells | Example volume | Score = weight × margin |
|---|
| buy cork yoga mat | Transactional (3) | 60% | 500 | 180 |
| best yoga mat for hot yoga | Commercial (2) | 45% | 2,400 | 90 |
| how to clean a yoga mat | Informational (1) | 45% (feeds product) | 12,000 | 45 |
| yoga mat | Informational (1) | 45% | 90,500 | 45 |
The head term has 180x the volume of the buyer term, yet scores a quarter of it — your first page slots go to the 500-volume buyer term. Chase revenue, not impressions.
Re-score quarterly, right after you re-run keyword research.