Hand Sprites a keyword list. It reads what Google actually ranks for each query, labels the intent and recommends the page format, so you stop publishing guides for buying terms.
What this template does
This template runs the SEO agent over a keyword list and answers one question per keyword: what does the searcher want, and what kind of page is Google rewarding for it? Sprites opens the live results for each query, looks at the pages that rank, the SERP features that appear (AI Overviews, People Also Ask, shopping units, video) and the wording of the query itself, then labels the intent as informational, commercial, transactional or navigational, with a note when the results are mixed.
The output is a table you can act on, not a score. Each row carries the intent label, the evidence Sprites saw, the format that ranks (guide, comparison, product page, tool, category page) and which of your own URLs already targets the query if Search Console is connected. Mismatches are called out: a blog post ranking on page two for a term where every top result is a product page is a rewrite, not a backlink problem.
From here the same agent can build topic clusters, write briefs, or draft and publish the pages through your WordPress, Shopify, Webflow or custom site integration, with every publish staged for your approval.
Inputs and outputs
What you provide
A keyword list: pasted, uploaded as CSV or spreadsheet, or linked from Google Drive or OneDrive
Optional: no list at all. Ask Sprites to research keywords first and it pulls volumes and competition through the Google Ads API
Your domain, so existing pages can be matched to each query
Optional: Search Console connected, so current impressions and positions sit beside each intent label
What you get
Intent label per keyword (informational, commercial, transactional, navigational, or mixed) with the SERP evidence behind it
Recommended page format per keyword, plus the angle the top results share and the one they miss
A map of your existing URLs to the queries they should own, with format mismatches flagged
Grouped keywords that share an intent and can be served by one page instead of several
A spreadsheet export and, if you want, a follow-on brief for any row
How it runs
Step 1
Load the keywords
Sprites reads your list, deduplicates close variants and, if you started without one, expands seed terms through Google Ads keyword data and filters out queries with no demand.
Step 2
Read the live results
For each query it checks what ranks today: page types, titles, SERP features and whether an AI Overview answers the question outright. Query modifiers such as "best", "vs", "pricing" and "how to" are weighed alongside the results, not instead of them.
Step 3
Label and group
Each keyword gets an intent label and a page-format recommendation. Queries that share intent are grouped so you plan one page per group, not one page per keyword.
Step 4
Match to your site
With your domain and Search Console connected, Sprites maps each group to the URL that should rank for it and flags gaps, cannibalization and format mismatches.
Step 5
Hand off
You get the table in chat and as a spreadsheet. Pick a row and Sprites writes the brief, drafts the page and stages it for publishing through your CMS integration.
Where teams use it
Content planning
Turn a raw keyword export into a plan with one page per intent group and a format for each, before anyone opens a document.
Fixing pages that stall
Find the posts that rank on page two because they answer a buying query with a tutorial, and rewrite them into the format that actually ranks.
Ecommerce category vs blog
Decide which product and category pages should own transactional terms and which questions belong in the guide section, so the two stop competing.
Agencies
Run the same mapping across every client workspace and hand each client a white-label intent report with the pages you plan to build.
By reading the live results, not just the words. It looks at the page types Google ranks, the SERP features shown and the query modifiers, then labels the intent and keeps the evidence in the row so you can check the call yourself.
Can it handle long-tail keywords and mixed intent?
+
Yes. Long-tail queries are usually the clearest to classify. When results mix formats, Sprites marks the keyword as mixed, names the dominant format and suggests whether one page can cover both or two are needed.
Do I need my own keyword list?
+
No. Start from a topic and Sprites researches keywords first through the Google Ads API, then maps intent for everything it found. If you already have an export from another tool, upload it and it is used as-is.
Does matching intent actually improve rankings?
+
It removes the most common reason a decent page does not rank: it is the wrong kind of page. Google ranks the format searchers engage with, so matching it is the prerequisite; quality, links and technical health still decide the rest.
Can I run this from Claude or ChatGPT?
+
Yes. Sprites has an official MCP server, so you can paste the keyword list into Claude, ChatGPT, Grok or any MCP-compatible assistant and get the same intent map, with the same integrations behind it.
Is this something my SEO agency would normally do?
+
Yes, and it is usually a slow deliverable. Sprites does the mapping in one run and keeps going into briefs, drafts and publishing, which is why teams use it to bring the work in-house instead of paying a retainer for it.
Related templates
Other prompts that run on the same agent. Browse the full gallery at /templates.
One template is a start. Sprites runs the whole account.
Every template here is one prompt to the same agent that plans, launches and optimizes campaigns across Google, Meta, LinkedIn, TikTok and Reddit, and keeps your SEO moving. Set a goal, approve the plan, and let it run.