SEO and content: Prompt library

SEO and content

Build intent-based keyword clusters from a few seed keywords

A ranked list of keyword clusters with volume, difficulty and the right page type or ad group for each.

I sell [product or service] in [city or country]. Start from these seed keywords: [3 to 5 seed keywords].

First check seo_history so you reuse earlier results instead of paying for the same query twice. Then use Keyword Planner and SEO keyword research to collect ideas with average monthly searches, competition and keyword difficulty.

Clean the list:
- remove keywords under 30 monthly searches
- remove job, free, DIY, course and competitor-brand terms unless they clearly fit my business
- merge close variants and plurals

Cluster the rest by search intent, with one cluster per distinct need a searcher has. For each cluster give:
- cluster name and head keyword
- 5 to 10 supporting keywords
- combined monthly volume and average difficulty
- intent: informational, commercial or transactional
- best use: service page, blog post or a Google Ads ad group

Rank the clusters by volume and fit for a small business with a limited budget. Mark the top 3 to start with and explain why in one line each.

SEO and content

Turn search terms that convert in ads into organic content clusters

Uses 90 days of converting Google Ads and Microsoft Ads search terms to find topics where you pay for conversions but have weak or no organic visibility. Returns a prioritized content plan and a preview-only list of keyword additions.

I want to use my ad data to decide what to write about. Use the last 90 days on every source.

1. Pull my Google Ads search terms for the last 90 days (days=90, or an explicit start and end date), sorted by conversions. Keep only terms with at least 1 conversion. If the response is truncated, page through until you have every converting term.
2. If Microsoft Ads is connected, find the default account and pull its search terms for the same 90 days. Keep only terms with at least 1 conversion and add them to the pool, marking which platform each term came from.
3. Group all converting terms into topic clusters by search intent. For each cluster show total cost, conversions and cost per conversion, split by platform where both are present.
4. Find my Search Console property (use [site property] if I named one) and pull organic data for the last 90 days with dimension query, plus a page,query pull so you can see which page ranks for each query. Use a high enough limit to cover the long tail.
5. Match each cluster to the Search Console queries that express the same intent. Use your judgement for close variants and wording differences, and say which queries you counted. Report organic clicks, impressions and average position per cluster, and the page that ranks best for it if there is one.

Classify each cluster:
- Content gap: converts in ads, organic average position worse than 10 or no organic impressions. I should build a page.
- Improve: converts in ads, organic position 4 to 10. Name the existing page that needs strengthening.
- Covered: organic position 1 to 3. Note it as a possible place to test lower ad spend later.

Deliver a table sorted by ad conversions, then the top 5 content-gap clusters with a suggested page title and URL slug for each.

Separately, list my Google Ads keywords (list_entities, entity keywords) and show converting Google search terms that are not yet keywords in the account. For Microsoft terms, just flag them; do not check or change Microsoft keywords. Do not add anything. If you think some Google terms should be added as exact match, tell me which ad group each one belongs in and run add_keywords with confirm=false so I see the exact preview. Wait for my approval before applying.

SEO and content

Cluster the searches that find my Business Profile and fix my services

Clusters of the real Google Search and Maps queries behind your listing, compared with your services and categories, plus a previewed fix.

For my Business Profile location [location name], pull the search terms people used on Google Search and Maps to find the listing over the last 3 months.

Cluster the terms into:
- brand searches
- service searches, with one cluster per service
- "near me" and neighborhood or district searches
- problem-based searches, such as a symptom or an issue the customer has

For each cluster show total impressions and the top 5 terms.

Then read my listing details: primary and additional categories, and the services list. Compare them with the clusters and tell me:
- service clusters with real demand that are missing from my services list
- services I list that get almost no searches
- whether an additional category would match a strong cluster. Search the category names to confirm they exist.

Draft the new service entries, with a name and a description of up to 300 characters for each, written around the words people actually search.

Show me the exact changes as a preview first. Do not update the listing until I approve.

SEO and content

Build a seasonal content and ad calendar from keyword demand

A 12-month calendar showing when each keyword cluster peaks, when to publish content and when to raise or lower ad budgets.

SEO and content

Rewrite meta titles for high-impression, low-CTR pages

A previewed set of new titles and descriptions for the 10 pages that waste the most first-page impressions.

Use Search Console property [client property] for the last 28 days. Find pages with at least 1,000 impressions, an average position of 1 to 10 and a CTR below 2%. Keep the 10 pages with the most impressions.

For each page, pull the top 3 queries by impressions. Then read its current meta title and description on WordPress site [WordPress site].

Deliver one table with these columns:
- URL, impressions, CTR, position, top queries
- current title and current description
- proposed title (max 60 characters, main query near the start)
- proposed description (max 155 characters, one concrete benefit or proof point, no clickbait)

Flag any page where the top query does not match the page's intent. Those pages need a content fix, not a new title.

Do not change anything yet. Show me a preview of every SEO field change and wait for my approval page by page. After I approve and the changes are applied, purge the cache. Then list the changed URLs with today's date so I can run a 28-day before/after CTR check.

SEO and content

Sitemap health check across all client properties

One agency-wide table of every property's sitemaps with errors, stale submissions and missing sitemaps flagged.

SEO and content

Paid vs organic overlap: where ads duplicate strong rankings

Matches Google Ads search terms with Search Console queries, separates brand traffic, classifies every overlap with non-overlapping rules, and finds high-impression organic queries with no paid coverage.

Compare Google Ads account [Google Ads account] with Search Console property [client property] for the last 30 days. Our brand terms are: [brand terms, comma separated, including common misspellings].

1. Pull the Google Ads search terms report for the last 30 days with min_clicks=0, so terms with impressions but no clicks are included. Pull Search Console queries for the same 30 days (dimension=query), sorted by impressions, with a limit high enough to cover the long tail.
2. Match Ads search terms to Search Console queries. Use exact text first, then close variants where the meaning is identical (singular and plural, word order, Turkish characters vs plain ASCII such as "ayakkabi" and "ayakkabı"). Mark each match as exact or variant.
3. Classify each matched query in this order. The first rule that fits wins, so every query gets exactly one class:
- BRAND: contains any of the brand terms above. List separately, no recommendation. Brand bidding is a client decision.
- TEST REDUCING: organic position 2.0 or better, organic CTR 20% or higher, and paid cost per conversion above [target CPA]. A query with paid cost above [target CPA] and zero conversions counts as above target.
- KEEP PAID: paid cost per conversion at or below [target CPA], or organic position worse than 5.
- MONITOR: everything else (for example organic position between 2.0 and 5 with paid cost per conversion above target, or a top-2 position with organic CTR under 20%). For each, say in one line what would move it into another class.

Separately, list non-brand queries with at least 500 organic impressions, position worse than 10, and no paid coverage. No paid coverage means the query does not appear in the search terms report and does not exist as a keyword in the account (check with list_entities, entity=keywords). These are candidates to add as paid keywords.

Deliver a table with these columns: query, match (exact or variant), organic position, organic CTR, organic clicks, paid cost, paid conversions, paid cost per conversion, class.

For TEST REDUCING queries, suggest a controlled test rather than removal: what to change, for how long, and the success metric (combined paid plus organic clicks and conversions for that query). Do not change bids, pause or remove anything.

If you propose adding keywords, list the ad groups with list_entities (entity=ad_groups), then show the exact keywords, match types and target ad groups by running add_keywords with confirm=false as a preview. Wait for my approval before calling it with confirm=true.

SEO and content

Page-two quick wins ranked by click potential

Finds Search Console queries stuck in positions 8 to 20, estimates the extra clicks from reaching position 4, and audits the ranking WordPress page for title, H1, intro, intent and internal links.

For Search Console property [client property], use the last 90 days. Pull query and page together (dimension=page,query) so each query comes with the page that ranks for it. Keep queries with an average position between 8 and 20 and at least 500 impressions. If two pages rank for the same query, keep the one with more impressions and note the other.

Estimate the extra monthly clicks if the query reached position 4: (impressions ÷ 3 × 7%) minus (current clicks ÷ 3). The 7% CTR at position 4 is a deliberately conservative assumption. Rank by this estimate and keep the top 15.

For the top 15, find the ranking page on WordPress site [WordPress site] (wordpress_list_posts with a search on the slug or title) and read it in full with wordpress_get_post. Read its SEO title with wordpress_seo, and use site_read_url with outline=true on the live URL to see the rendered title and H1. Then check:
- whether the query, or a close variant, appears in the SEO title, the H1 and the first 100 words of the content
- whether the content covers the intent behind the query (your judgement, with a one-line reason)
- internal links: search the site content for the page's path with wordpress_search_content to find which posts and pages link to it. Cross that list with the pages that rank in the top 5 for any query in the same Search Console data. Report how many top-5 pages link to it, and flag it if fewer than 2. Note that this only sees links inside post content, not menus, widgets or theme templates.

Deliver a table with these columns: query, page, position, impressions, estimated extra clicks, gaps found, recommended action. Recommended action is one of: retitle, expand section, add internal links, or new dedicated page. For "add internal links", name the top-5 pages that should link and the anchor text to use.

Do not edit any content. If I ask you to apply a change, show me the exact before and after text first and wait for my approval.

SEO and content

Diagnose an organic traffic drop: rankings, CTR or demand

A short diagnosis that pins the drop to specific pages, queries and causes, backed by Search Console and GA4 evidence.

Client [client name] reports an organic traffic drop. Use Search Console property [client property]. Compare the last 28 days with the previous 28 days.

1. Break down clicks by page, query, device and country. Show which 10 pages and 10 queries lost the most clicks.
2. For each losing page, split the loss into three causes:
   - fewer impressions at a stable position: demand or seasonality
   - a position decline: rankings
   - a stable position with lower CTR: a SERP change or the snippet
3. Inspect the top 10 losing URLs. Report index status, last crawl date and canonical, and flag anything not indexed or canonicalised elsewhere.
4. Check sitemaps for new errors.
5. In GA4, compare organic sessions by landing page for the same periods. This confirms whether the drop is real on the site and not only in Search Console.

Deliver a one-page diagnosis:
- the headline cause, with numbers
- a table of the losing pages with cause type
- the 3 most likely explanations, ranked, each with its evidence and the next check or fix

Read-only. Do not change anything.

SEO and content

Post-redesign check: old URLs with traffic that now break

Rebuilds the pre-launch page list from Search Console, checks each old URL live, maps broken ones to the closest new page by topic and intent, and creates 301 redirects on the hosting server only after approval.

Client site [client domain] was relaunched on our hosting server on [launch date]. Use Search Console property [client property]. If the server hosts more than one site, the site key is [site key].

1. Pull every page that had clicks or impressions in the 90 days before the launch date. Use get_report with platform=search_console, start_date and end_date set to that window, at page level (check describe for search_console if you are unsure of the level name). If that window cannot be read, fall back to gsc_performance with dimension=page and enough days to reach back before the launch, and tell me the totals then include post-launch data. Search Console keeps about 16 months, so say so if the window is older than that.
2. Fetch each of these URLs live with site_read_url and record the HTTP status code and, for redirects, the Location header.
3. Read the redirects that already exist with site_redirects (mode=list).
4. To understand the new site, list its pages with site_list_files and read the relevant ones with site_read_file, or with site_read_url outline=true for title and headings.

Deliver a table with these columns: old URL, clicks, impressions, current status (200, 301, 404 or other), existing redirect target, proposed new URL.

To choose the proposed new URL, match on topic and intent, not just slug similarity. Point to the closest equivalent page, never blanket-redirect to the homepage. Mark any URL with no sensible equivalent as "let 404". Also flag redirect chains (an old URL whose target redirects again) and redirects that land on a non-200 page, and propose the direct fix.

Show me the full redirect map, then run site_redirects with mode=add and confirm=false so I see the exact preview. Wait for my approval before calling it with confirm=true.

After I approve and the redirects go live:
- re-fetch each old URL with site_read_url and confirm it returns a 301 to the right target
- inspect the target URLs with gsc_inspect_urls in batches of up to 20, highest-traffic old URLs first. If the quota runs out, tell me which ones are left so we continue later
- read the current sitemaps with gsc_sitemaps, propose resubmitting the right one, and wait for my confirmation before calling gsc_submit_sitemap

SEO and content

Commercial keyword gap vs competitors, mapped to our category pages

Runs a keyword gap against each competitor, keeps only non-brand commercial terms, checks our current rankings, maps each keyword to an existing category or product page and prioritises them with a clear A/B/C rule.

Our domain is [our domain], competitors are [competitor domains, comma separated], market is [country] in [language]. Our brand terms: [our brand terms]. Competitor brand terms: [competitor brand terms]. Our category and product pages are listed in [sitemap URL or category list].

1. Check seo_history first and reuse any recent run with the same arguments for free. Then run seo_competitor_gap for [our domain] against each competitor separately, and seo_ranked_keywords for our domain. Tell me roughly how many SEO Intelligence credits this will use before you start.
2. Keep only commercial and transactional keywords: product names, categories, "buy", "price", "discount", brand plus model. Drop informational queries, and drop any keyword containing our brand terms or competitor brand terms.
3. For each remaining keyword, check whether we already rank: use our seo_ranked_keywords data and Search Console property [client property] for the last 90 days with dimension=page,query, which also tells you which of our pages ranks.
4. For keywords where we have no ranking page, read the page list with site_read_url (or use the list I pasted) and pick the closest existing category or product page, or say "new page needed".

Deliver one table with these columns:
- keyword
- monthly volume
- keyword difficulty (if the gap data includes it, otherwise leave blank)
- best competitor position, and how many competitors rank in the top 10 (count across the separate gap runs)
- our position (or "none")
- suggested target URL (an existing category or product page, or "new page needed")
- priority

Decision rule, in this order:
- Priority A: volume ≥ 500, at least two competitors in the top 10, and we have a page in positions 11-50.
- Priority B: we have no ranking page yet, but there is a clear fit with our catalog.
- Priority C: everything else.

Finish with:
1. The top 15 A keywords, grouped by target URL, so I can see which pages to work on first.
2. A proposed list of 20 keywords to add to the rank tracker, with the location, language, device and depth you would use.

Do not run seo_rank_tracker until I approve that list. It costs credits and keeps re-measuring daily while auto is on.

SEO and content

Competitor paid keywords filtered by our break-even CPC

Takes a competitor's paid search keywords and ad copy, removes what you already cover, and keeps only terms whose top-of-page bid fits under a break-even CPC calculated from your own conversion rate, AOV and target ROAS. It then previews the keyword additions.

Find the keywords [competitor domain] pays for that we could profitably add to [account]. Market: [country], language: [language].

Step 1: competitor data.
- Check seo_history first and reuse any stored result.
- Pull the competitor's paid keywords, with their ad headlines, descriptions and landing pages, for our market.
- Also check their live ads in the Google Ads Transparency Center. If that returns little or nothing, rely on the ad copy from the paid keyword data and say so.

Step 2: filter.
- Keep only product and category terms that match our catalog: [main categories].
- Remove anything that is already a keyword in [account], active or paused (keyword report, including zero-impression keywords).
- Remove anything that appeared in our search terms in the last 90 days.

Step 3: market data. Find the geo id for [country], then get Keyword Planner historical metrics for what remains with that geo and the [language] language id: average monthly searches and the top-of-page bid range.

Step 4: our break-even CPC. Run a google_ads_query on the campaign resource for the last 90 days (clicks, conversions, conversions_value) for the campaigns that sell [main categories], and calculate:
- conversion rate = conversions / clicks
- AOV = conversion value / conversions
- break-even CPC = AOV × conversion rate / [target ROAS]
If those campaigns have fewer than 30 conversions, say the number is unreliable and calculate it at account level as well.

Deliver a table with these columns:
- keyword
- monthly searches
- high top-of-page bid
- break-even CPC
- competitor ad message (from their ad copy, if found)
- verdict

Verdict rule: "add" only when the high top-of-page bid is at or below break-even CPC. Otherwise "skip".

For each "add" keyword, list our ad groups and propose the best-fitting target ad group, with one line explaining why. Use phrase match by default, and exact match for keywords with fewer than 300 monthly searches. Show the full add_keywords preview (confirm=false) per ad group and wait for my approval.

SEO and content

Page-2 product and category queries ranked by conversion potential

Finds Search Console queries at positions 8 to 20 on product and category pages, scores them by impressions times the page's organic conversion rate from GA4, drops marketplace-dominated SERPs, and previews new SEO titles and meta descriptions in WordPress.

Find page-2 quick wins on pages that actually convert. Site: [site], GA4 property: [GA4 property].

Step 1: candidate queries from Search Console, last 90 days.
- Pull page,query rows with a high limit.
- Keep rows with average position 8 to 20 and at least [minimum impressions] impressions.
- Keep only product and category URLs: [product URL pattern, e.g. /urun/] and [category URL pattern, e.g. /kategori/]. Exclude /blog/ and other informational paths.

Step 2: conversion data from GA4.
- List the key events and confirm the purchase event name.
- For the same 90 days, get organic sessions and purchases per landing page (landing page with channel as the second dimension, plus landing page × event if there are several key events). Page conversion rate = organic purchases / organic sessions.
- Score each query: impressions × page conversion rate. Pages with no organic purchases score 0; list them separately, do not drop them silently.

Step 3: live SERP for the top 15 by score.
- Check seo_history first and reuse stored SERPs.
- Count how many of the top 10 organic results belong to these marketplaces and large retailers: [marketplace and big retailer domains, e.g. trendyol.com, hepsiburada.com, amazon.com.tr, n11.com]. If 5 or more, mark the query "hard" and drop it.

Step 4: current SEO fields.
- Product and page URLs: find the post id by slug with wordpress_list_posts (type product for WooCommerce products), then read its SEO title and meta description.
- Category URLs are archives, not posts: find the term id with wordpress_terms and read the archive's SEO fields by term id.

Deliver the top 10 remaining queries in a table with these columns:
- query
- page
- position
- impressions
- click-through rate
- page conversion rate
- score
- current title and meta description
- proposed title (≤ 60 characters)
- proposed meta description (≤ 155 characters)

The proposed title must include the query. Do not change a title that already contains it, unless its click-through rate is below 2%. If two queries point to the same page, write one title that serves the higher-scoring query.

Show the WordPress SEO changes as a preview (page id or term id, old value, new value) and wait for my approval before writing anything. After I approve, write them with wordpress_set_seo and purge the cache.

SEO and content

Outreach targets from competitor link intersect

A ranked list of 25 clean sites that link to your competitors but not to you, each with an outreach angle and two email templates.

SEO and content

Anchor text mix check for local SEO

Your anchor text split into branded, URL, generic, exact, partial and local groups, compared with a competitor, plus a target mix for the next links.

SEO and content

Pass link equity from top-linked pages to service pages

Finds the pages that already earn backlinks, matches them to service pages stuck in positions 5 to 20, and proposes contextual internal links. Each edit is previewed and applied only after approval.

Use the pages that already earn backlinks on [my domain] to push my service pages up.

My service pages are: [list of service page URLs, or a URL pattern such as /services/]. Treat only these as targets. If I left this empty, propose a list based on URLs and page titles and wait for my confirmation before going on.

1. Check seo_history first. If a best-pages-by-links run for [my domain] is still reusable, use it so I do not pay twice. Otherwise pull my best pages by backlinks and keep the top 10 by referring domains.
2. From Search Console (use gsc_list_sites to pick the right property), pull page-level data for the last 90 days sorted by impressions. Keep the service pages with an average position between 5 and 20 and at least 200 impressions. These are the targets.
3. Find the WordPress post ID of every source and target page with wordpress_list_posts (search by slug or title). Show me the URL to ID mapping, and list any URL you could not match.
4. Read each top-linked source page in full with wordpress_get_post. Also run wordpress_page_design on it so we know whether it is built with Elementor, WPBakery, Gutenberg or the classic editor.

For each source page, propose 1 or 2 contextual internal links to the most relevant target page. Show: source page, target page, anchor text (descriptive, not stuffed with keywords), and the exact sentence where the link goes, as before and after.

Rules:
- No more than 2 new links per source page.
- Do not link from a page to a service it does not relate to.
- Prefer existing sentences. Only add a new sentence if no natural spot exists.
- Skip any source page that already links to that target, and say so.

This is a preview. Do not update any page until I approve each edit. After approval:
- On classic, Gutenberg or WPBakery pages, use wordpress_update_post with a find/replace on the exact sentence, never a full content overwrite.
- On Elementor pages, use wordpress_edit_page_design on the right text element.
- Then purge the page cache, re-read each edited page and run wordpress_search_content for each target URL to confirm the new link is in the stored content. Report every edit as confirmed or failed.

SEO and content

Which backlinks actually send leads

Joins GA4 referral traffic and key events with your live backlink list to show which linking sites send leads, which send empty traffic and which are SEO-only, then suggests where to look for similar links.

Show me which backlinks to [my domain] bring real leads, not just link value.

1. List my GA4 key events first. Confirm that phone clicks and form submissions are tracked as key events. If they are not, stop and tell me what is missing before going further.
2. Pull GA4 sessions and key events by source / medium for the last 180 days. Keep referral traffic only and exclude payment gateways, my own domains and email providers. If the report can split by event name as a second dimension, show phone and form key events separately; otherwise use total key events and say so.
3. Check seo_history, then pull my live backlink list (all link types) for [my domain]. Group it by linking domain.
4. Match each GA4 referral source to a linking domain. Normalise subdomains (m., www., l.) to the root domain. List referral sources with no matching backlink separately, and mark every match you are unsure about.

Deliver a table with columns: referring domain, sessions, key events, conversion rate, link type (dofollow / nofollow, from the backlink list), and the page of mine the link points to.

Then sort the domains into three groups:
- Lead sources: 2 or more key events.
- Traffic, no leads: 50 or more sessions, zero key events. Check whether the linked page from the backlink list is the right page for that audience.
- SEO-only: linked but no meaningful traffic.

To find more sites like my lead sources:
- Run a link intersect on these competitors, excluding [my domain]: [2-3 competitor domains]. Keep domains with low spam scores.
- Run a live SERP analysis for "[main service] [city]" and note which directories, local media or association sites rank.

End with 5 concrete site types in [city] that resemble my lead sources, each with example domains from the intersect or SERP data. Mark clearly which suggestions come from data and which are your judgement.

SEO and content

Build a management-ready keyword tracking set

Combines Search Console queries, ranked keywords and your must-report terms into a deduplicated list of up to 30 keywords, checks for reusable paid data, and starts daily rank tracking only after approval.

Help me build the keyword set I will report to management every month for [domain]. Our brand name is [brand name]. Market: [country], language: [language], device: [mobile or desktop].

1. Pull Search Console queries for the last 90 days, sorted by impressions. List the top 40 with clicks, CTR and average position. Pull the page and query combination as well, so each query has a ranking URL.
2. Check seo_history and seo_rank_history first. Reuse any stored ranked-keywords or research run that is still free, and tell me which keywords are already being tracked.
3. Pull our ranked keywords from SEO Intelligence. Add any non-brand term with monthly search volume above [minimum search volume] that is missing from the Search Console list.
4. Merge in these business-critical terms I must report on regardless of data: [priority keywords]. If any of them has no search volume yet, use keyword research to fill it in.
5. Propose a final list of no more than 30 keywords, grouped into Brand, Core product, Category and Informational. Use one row per keyword with these columns: keyword, group, current position, ranking URL, search volume, and a one-line reason it belongs.

Remove near-duplicates such as singular/plural forms or reordered words, and keep the variant with the most impressions.

Before tracking anything, show me:
- The final list, marking which keywords are already tracked and which are new.
- What the tracking will cost in credit terms: number of new keywords, scan depth (keep the default 20 unless I say otherwise) and the fact that daily re-measurement keeps using credits. You do not need an exact price; I will check my add-on balance in the panel.

Start daily rank tracking for the new keywords only after I approve.

SEO and content

Monthly organic ranking report for management

A read-only monthly update built from stored rank history, Search Console and GA4: position buckets, top winners and losers, organic sessions and key events, and a plain-language executive summary.

Prepare the monthly organic ranking update for management for [domain]. Compare the last 30 days with the 30 days before that. I run this in the first days of each month.

1. Read stored rank history for our tracked keywords over the last 60 days. It is free and needs no new credits. If the history does not cover both periods (tracking started recently or was stopped), say so clearly, and use Search Console average position for the missing period instead.
2. From Search Console, pull query and page data for the last 30 days and for the last 60 days. Derive the previous 30 days' clicks and impressions by subtracting the first from the second. Do not subtract positions; take positions from rank history.
3. From GA4, pull sessions and key events by landing page for the last 30 and the last 60 days, split by channel, and keep Organic Search only. Derive the previous period the same way.

Deliver the following:
1. A bucket table showing the count of keywords in the top 3, top 10, positions 11-20 and not ranking, for this period vs the previous one.
2. The top 5 winners and top 5 losers by position change. Include only moves of 3 or more positions. Show the ranking URL, the change in clicks and the change in impressions.
3. Organic sessions and key events for the landing pages behind those winners and losers.
4. A 5-bullet executive summary in plain business language covering what moved, the likely reason and what we do next. If you use an SEO term, explain it in one line.

Flag every keyword whose position improved while clicks fell. That pattern often means a SERP feature or an AI Overview is taking the click, but treat it as a hypothesis. Check seo_history for any AI Overview check already run on those keywords and use it. If I want confirmation for the rest, list them and ask before running new AI Overview checks, since they cost credits.

This is reporting only. Do not change anything.

SEO and content

Diagnose ranking drops before reporting them

A read-only diagnosis of tracked keywords that lost 3 or more positions: rank history, click loss, live SERPs, index status and on-page SEO fields, ending in a cause and fix table.

Several tracked keywords on [domain] have lost positions. I need a diagnosis before I report this upward. Market: [country], language: [language].

1. From rank history, list every tracked keyword that dropped 3 or more positions in the last [number of days] days. Show the old position, the new position and the ranking URL. If the history shows a different ranking URL over the period, flag it, because that signals cannibalisation. If the history does not carry URLs, check Search Console page and query data instead: two of my pages taking impressions for the same query is the same signal.
2. From Search Console, pull clicks for those queries over the last [number of days] days and over twice that period, and derive the previous period by subtracting. Take the 5 drops that lost the most clicks.
3. Check seo_history for SERP analyses already run on these keywords and reuse them. For the rest, tell me how many new SERP analyses you need, since each costs add-on credits, and wait for my go-ahead. Then run them: who now ranks above us, and which SERP features appear.
4. Inspect those ranking URLs in Search Console in one batch. Report the index status, the last crawl date and the Google-chosen canonical.
5. Check whether [domain] is a connected WordPress site. If it is, find each page's post ID with wordpress_list_posts and read its SEO fields: meta title, meta description and canonical. If it is not, skip this step and say so.

Output a table with these columns: keyword, drop size, clicks lost, likely cause, evidence, recommended fix, owner (content or dev). Choose the likely cause from: deindexed, canonical conflict, URL swap, new competitor, SERP feature, unclear. Use "unclear" whenever the evidence does not point to one cause.

This is diagnosis only. Do not change anything.

SEO and content

Are AI Overviews eating our clicks on tracked keywords?

Checks your tracked keywords for Google AI Overviews and citations, compares Search Console CTR over the last 90 days vs the prior 90, and flags the pages losing clicks to AI answers, with fixes and a management summary.

Management is asking whether Google AI Overviews are hurting our organic traffic. Answer this for [domain], brand name [brand], market [country], language [language code, e.g. en or tr].

1. Get the keyword list. Read the stored rank history for [domain] over the last 90 days. This is free and gives each tracked keyword with its latest position. If nothing is tracked, use this list instead: [paste keywords, one per line].
2. Find the Search Console property for [domain] with gsc_list_sites. Then pull performance by query twice, once for the last 90 days and once for the last 180 days. Derive the prior 90 days from these two pulls:
   - clicks and impressions = 180-day total minus 90-day total
   - CTR = derived clicks ÷ derived impressions
   - position ≈ impressions-weighted difference of the two averages (label it as an estimate)
   Also pull page,query for the last 90 days so you know which URL ranks for each keyword.
3. Before spending credits, check seo_history and reuse any stored AI Overview or citation results that are still reusable. The AI Overview check costs SEO add-on credits and needs one call per keyword. Tell me how many new calls are needed. If there are more than [10], propose the top [10] by Search Console impressions. Wait for my go-ahead before running them.
4. For each approved keyword, run the AI Overview check with [domain] as the target. Record whether an AI Overview appears, whether we are cited, and which domains are cited instead of us.
5. Run AI-answer citation tracking once for [brand] as an overall baseline of how often we are mentioned in AI Overviews and ChatGPT.

Output a table with these columns: keyword, organic position, AI Overview shown (yes/no), we are cited (yes/no), cited competitors, CTR last 90 days, CTR prior 90 days, CTR change.

Then list the priority keywords that meet all three conditions:
   - an AI Overview is shown
   - we rank in the top 5 but are not cited
   - CTR fell more than 20%

For each priority keyword, name the ranking URL. Using what the AI Overview text says and what the cited sources cover, suggest what our page lacks: for example a direct answer, a definition, a comparison table or updated facts.

Close with 4 plain-language bullets for management, and state clearly which numbers are estimates.

SEO and content

Push page-two keywords to page one with better snippets

Finds queries ranking in positions 4 to 15 whose CTR lags the site's own benchmark, filters out unstable rankings, and rewrites WordPress meta titles and descriptions after row-by-row approval.

Find quick ranking wins on [domain]: queries already close to page one where a better title and description can earn more clicks.

1. Run gsc_list_sites and pick the property for [domain]. Run list_wordpress_sites and pick the matching WordPress site.
2. Pull Search Console data for the last 28 days with dimension page,query. Keep the query and page pairs with an average position between 4 and 15 and at least [minimum impressions, e.g. 100] impressions.
3. Build the CTR benchmark from the site's own data, since no tool provides an industry benchmark. Group all page,query rows that meet the impression minimum into position bands 4-6, 7-10 and 11-15, and calculate the median CTR of each band. Flag pairs whose CTR is at least 30% below their band's median. Show me the band medians so I can see the reference.
4. Check stability. If a query is tracked, read its point series with seo_rank_history (target [domain], days 30). If it is not tracked, pull gsc_performance with dimension date, filtered to that query and its page, for the last 30 days. Drop any keyword whose position moved more than 5 places in that window, and note which source you used for each.
5. Take the top 10 remaining pairs by impressions, flagged ones first. Match each ranking URL to its WordPress ID: search wordpress_list_posts (types post and page) using words from the slug and confirm the returned link equals the URL. If the URL is a category or tag archive, find its term_id with wordpress_terms. If a URL cannot be matched, say so and leave it out.
6. Read the current meta title and description with wordpress_seo. Note fields that are empty and fall back to the plugin's template.
7. Draft a new title of up to 60 characters and a new description of up to 155 characters for each. Use the query naturally, state a concrete benefit and avoid clickbait.

Show a before/after table with these columns: URL, query, position, impressions, CTR, band median, current title, new title (character count), current description, new description (character count).

Do not apply anything yet; I will approve row by row. After I approve, call wordpress_set_seo only for the approved rows (id for pages and posts, term_id for archives, overwrite=true where a value already exists). Read each one back with wordpress_seo to confirm it saved, then run wordpress_purge_cache once. Check one updated page with site_read_url (outline=true) to confirm the live title changed.

Finally give me a table with each URL, today's date as the change date, and its baseline position and CTR, so in 4 weeks I can rerun gsc_performance for the 28 days after the change date and compare.

SEO and content

Clean up the rank tracking list to cut credit spend

Audits every tracked keyword against stored rank history and Search Console demand, sorts them into Keep, Merge and Drop, and untracks only the approved ones while keeping their history.

Our rank tracking list for [domain] has grown over time, and every tracked keyword is re-measured daily at a credit cost. Audit it.

Inputs I will paste below, since no tool holds them:
- Keywords in our management report: [paste list]
- Discontinued products or services: [paste list]
- Topics with content planned: [paste list]
- Optional: our daily credit cost per tracked keyword: [amount]

1. Read seo_rank_history for target [domain] with days 90 (this is free). For every tracked keyword, list the first measured date, current position, best position and the position range over the last 30 days. Treat "position null, measured true" as outside the scanned depth, not missing data.
2. Run gsc_list_sites to pick the property for [domain], then pull gsc_performance with dimension query for the last 90 days, sorted by impressions, with a limit high enough to cover the tracked list. Match each tracked keyword to its clicks and impressions. If a keyword is not in the rows, query it directly with the query filter before marking it as zero.
3. Classify each keyword:
- Keep: it is on my management report list, or it had more than 100 impressions.
- Merge: it is a near-duplicate of another tracked keyword (word order, plural, accents) and ranks with the same URL. Keep the variant with more impressions.
- Drop: zero impressions, off-topic for the site, tied to a discontinued product on my list, or outside the scanned depth for 60+ days with no matching topic on my content list.

Output a table with these columns: keyword, class, reason, impressions, clicks, current position, ranking URL.

Then summarise: how many keywords stay tracked and how many would stop. If I gave a per-keyword daily cost, multiply it out for daily and monthly savings; if not, give only the keyword counts and do not guess a credit figure.

On history: per the tool, seo_rank_untrack keeps stored history by default, so dropped keywords stay readable in seo_rank_history. Never pass delete_history=true.

Do not untrack anything until I approve the final Drop and Merge lists. After approval, call seo_rank_untrack once with exactly the approved keywords, then rerun seo_rank_history to confirm their history is still there.

SEO and content

Measure my share of voice in AI answers against my competitors

A share-of-voice table showing how often you and your competitors are mentioned or cited in AI answers to the questions your customers actually ask.

I want to know how visible my business is in AI answers compared with my competitors.

My business is [my domain], and I offer [service] in [city or region]. My competitors are [3 competitor domains].

1. Draft 10 questions a real customer would ask an AI assistant before buying. Cover four types:
   - "best [service] in [city]"
   - cost
   - comparison
   - how-to

Show me the list and wait for my approval before running anything, because these checks use add-on credits.

2. Before running, check which of these queries were already run recently. Reuse any stored results instead of paying again.

3. After I approve the list, run the share-of-voice and citation checks for my domain and the three competitors.

Deliver:
- A table with these columns: question, which brands are mentioned, which domains are cited, my position in the answer (or "absent").
- A summary row with each domain's share of voice in %.
- My 3 weakest questions, meaning the ones where competitors are cited and I am absent. For each, name the specific page type that gets cited, such as a pricing page, a guide or a directory listing.

Do not make any changes.

SEO and content

SERP-based content brief for a new target keyword

Checks Search Console first so you don't create a page that competes with one you already have, then builds a writer-ready brief from the live SERP, the competitor outline, People Also Ask and real keyword volumes.

Build a content brief for the keyword [keyword] in [country and language] for the client domain [domain].

Step 1. Cannibalisation check. Find the matching property for [domain] with gsc_list_sites. Then run gsc_performance for the last 90 days with dimension "page,query", filtered to the query [keyword]. If any page on [domain] gets impressions for this keyword at an average position of 20 or better, stop and recommend updating that page instead of creating a new one. Give the URL, its current average position, impressions and clicks. If several pages show up for the same query, flag it as cannibalisation and say which page should own the keyword.

Step 2. Before paying for any SEO data, check seo_history. If a recent seo_serp_analysis, seo_content_brief or seo_keyword_research run for this keyword and location is still reusable, call it again with the same arguments so it costs nothing.

Step 3. If no existing page qualifies, run seo_serp_analysis (depth 10), seo_content_brief and seo_keyword_research with [keyword] as the seed, all for [country and language]. Produce the brief with these parts:
- Dominant search intent and the content format that wins (guide, list, comparison, landing page), backed by what the top 10 actually shows: the domains, page titles and any paid results.
- Recommended title tag (max 60 characters), meta description (max 155 characters) and H1. Count the characters and show the count.
- Full H2/H3 outline, starting from the #1 competitor's outline in the brief data and improving on it. For each section list the subtopics, entities and People Also Ask questions it must cover.
- Target length and depth: base this on how many sections and questions the #1 competitor's outline covers and on the content format the SERP rewards. No tool measures the word count of ranking pages. If I want a hard word count range, I will paste the word counts of the top 5 URLs from my own crawler, and you should use those instead.
- 5 to 10 secondary keywords, taken from the keyword research and brief results, with the monthly search volume and difficulty those results return. Only include keywords that come back with real volume data, and label any you add without data as "no volume data".
- SERP features to target (featured snippet, People Also Ask, video, others), only the ones the brief data shows this query actually triggers, with one line on how to format the content to win each.
- 3 internal links to add: run gsc_performance for the last 90 days with dimension "page,query" and pick 3 existing [domain] pages that get impressions for queries closely related to [keyword]. For each, give the source URL, the related query it ranks for, its position and the suggested anchor text.

Keep it to what a writer needs. No generic SEO advice. After the brief is written as a draft, remind me to run seo_content_lint on it before publishing.

SEO and content

Rewrite titles and meta descriptions on pages with below-par organic CTR

Benchmarks each page's CTR against the site's own median for its position band, rewrites titles and descriptions for the weakest high-impression pages, and applies only approved values in WordPress.

For the Search Console property [property], find pages whose snippet is losing clicks.

1. Confirm the property with gsc_list_sites and the matching site with list_wordpress_sites.
2. Pull gsc_performance with dimension page for the last 28 days, sorted by impressions. Keep pages with at least 1,000 impressions and an average position of 10 or better.
3. Group them into position bands 1-3, 4-6 and 7-10 and calculate the site's median CTR for each band yourself from these rows. Show the medians and how many pages each band holds; if a band has fewer than 5 pages, say the median is weak.
4. Flag pages whose CTR is at least 30% below their band's median.
5. For the 10 flagged pages with the most impressions:
   - Find the WordPress ID by searching wordpress_list_posts (types post and page) with words from the slug, and confirm the link matches the URL.
   - Read the current meta title and description with wordpress_seo.
   - Pull gsc_performance with dimension query filtered to that page and list the top 3 queries by impressions.
6. Deliver a table with these columns: URL, position, CTR, band median, current title, current description, top 3 queries, and two rewrite options. Each title must be 60 characters or fewer and include the main query. Each description must be 155 characters or fewer, with a concrete benefit or proof point taken from the page itself.
7. Flag pages where Google is likely rewriting the title: the current title is over 60 characters, is empty, or does not match the intent of the top queries. If I approve the credits, check the 3 most important ones with seo_serp_analysis on their main query to see the title Google actually shows.

Do not change anything yet. After I choose an option per page, show the final values as a preview table. Wait for my confirmation, then apply them with wordpress_set_seo (overwrite=true where a value exists), read each back with wordpress_seo, run wordpress_purge_cache, and spot-check one page with site_read_url (outline=true).

Record today's date and each page's baseline CTR and position in a final table. There is no scheduler, so in 28 days I will rerun gsc_performance for the 28 days after that date and compare.

SEO and content

Turn converting paid search terms into organic content briefs

Finds search terms that convert profitably in Google Ads but have no organic top-10 page, clusters them by intent, and writes SERP-grounded content briefs for the top clusters. Read-only.

Use paid search data from the Google Ads account [account] to decide what organic content to build next. The site's Search Console property is [property] and the target market is [country, language].

1. Confirm the account with list_accounts and the property with gsc_list_sites. Run list_conversion_actions to see which actions are included in Conversions, so the CTA in step 6 matches the real goal.
2. Pull search_terms_report for the last 90 days, sorted by conversions. Keep terms with 2 or more conversions and a cost per conversion (cost divided by conversions) below [target CPA].
3. For each term, pull gsc_performance with dimension page,query and the query filter for the same 90 days. Show the organic position and ranking URL, or "not ranking" if there are no rows. Note when only a close variant ranks rather than the exact term.
4. Keep the terms with no page in the organic top 10. Group them into topic clusters by shared intent.
5. Get monthly search volume with keyword_historical_metrics (language_id and geo_ids for [country, language], e.g. 1037 for Turkish). If Keyword Planner data is not available on the account, tell me and ask before using seo_keyword_research, which costs add-on credits; otherwise leave the column as "n/a".
6. Deliver a table ranked by paid conversions with these columns: cluster, terms included, paid conversions, paid spend, cost per conversion, best organic position, monthly search volume.
7. Pull rsa_asset_performance for the campaigns behind these terms and note the headlines and descriptions with the best cost per conversion. These are the proven proof points.
8. For the top 3 clusters, run seo_content_brief on the cluster's main term (tell me first that it costs add-on credits). Then write a brief per cluster with:
   - the page type,
   - a title tag (max 60 characters) and an H1,
   - an H2/H3 outline,
   - the questions to answer, from the People Also Ask tree,
   - the proof points from step 7,
   - a primary CTA matching the conversion action from step 1.

This is analysis only. Do not change anything in the Google Ads account. Paid budgets on these terms stay as they are until the organic pages rank.

SEO and content

Pre-publish quality check for a client draft in WordPress

Reads a WordPress draft and its SEO fields, runs the AI-footprint linter, checks intent, headings, concrete detail, meta lengths and internal links, then gives a ready or not-ready verdict with rewritten paragraphs. Nothing is saved without approval.

Review the draft post [post ID or title] on the client site [site] before it goes live.

1. If I gave a title instead of an ID, find the post among the site's drafts. Read the full draft with its raw content. If the content looks empty or is only builder shortcodes, inspect the page design to read the real text. Then read its SEO fields: meta title, meta description and focus keyword. If no focus keyword is set, tell me and ask which one to use.
2. Run the AI-footprint linter on the body text. List every flagged phrase or pattern with a specific replacement. If the linter does not report paragraph numbers, map each flag to its paragraph yourself by finding the phrase in the text.
3. Check the draft on these points:
   - Do the post title (which the theme usually prints as the H1) and the first 100 words match the focus keyword's search intent?
   - Is the heading hierarchy in the content clean? There should be no extra H1 inside the body and no skipped levels (H2 straight to H4).
   - Are concrete details missing: numbers, examples, first-hand experience, a named author or expert?
   - Do the meta title and description fit 60 and 155 characters?
   - Does the body link to at least 3 relevant existing posts? List the site's published posts, searching by the draft's main topics, and suggest 3 to 5 specific posts with the anchor text and the paragraph where each link belongs.
4. Decision rule: count the paragraphs that carry at least one linter flag. If that is more than a third of all paragraphs, or the draft has no concrete examples at all, mark it "not ready" and rewrite every flagged paragraph in full. Otherwise mark it "ready after edits".

Deliver the verdict first, then the fix list, then the rewritten paragraphs side by side with the originals.

Do not save anything to WordPress until I approve. If I approve, first show me the exact text replacements and meta field values you will write, then apply only those. Keep the post in draft status. Never publish it.

SEO and content

Fix organic landing pages that get traffic but fail to engage

Finds high-traffic landing pages with weak engagement or zero key events in GA4, checks their Search Console queries and mobile Core Web Vitals, inspects how each WordPress page is built, diagnoses the main problem and previews exact page edits for approval.

For the GA4 property [GA4 property], the Search Console property [property] and the WordPress site [site], find landing pages that attract visitors but lose them.

1. Pull the site-wide GA4 overview for the last 28 days to get the average engagement rate. Then pull a landing page report for the same 28 days with sessions and key events. Keep pages with at least 300 sessions where either:
   - the engagement rate is 15 or more points below the site average (if the report does not give engagement rate per page, use key events per session against the site average instead and tell me you did), or
   - key events are zero.
   Cap the list at the 10 pages with the most sessions.
2. For each page, pull its top 5 Search Console queries for the last 28 days, filtered to that page URL, and judge whether the content answers what people searched for.
3. Run a mobile Core Web Vitals check on each page, one call per page. Note LCP, CLS and INP and the LCP element.
4. Match each URL to its WordPress page or post by slug. Inspect how it is built:
   - On an Elementor page, read the design elements: what sits in the first section, where the first button or CTA appears, and how far down the answer to the main query starts.
   - On other builders or the block editor, read the raw content and judge the same points from the block or shortcode order.
   If a URL does not map to a WordPress page, say so and skip step 4 for it.
5. Deliver one row per page with these columns: sessions, engagement rate (or key events per session), key events, top queries, main problem, and 3 specific fixes. The main problem must be one of: intent mismatch, buried answer, weak or missing CTA, slow LCP, or layout shift. Rank the rows by sessions.

For the top 3 pages, propose exact edits: heading text, button label, link, and block order. For each edit, show the element or block, the current value and the new value. Do not change anything until I approve each page. After I approve, apply only the approved edits (design edits on Elementor pages, text replacements in the content on other pages) and then purge the site cache. If a page looks wrong afterwards, roll it back to the previous version.

SEO and content

Triage unanswered Google reviews and draft replies for approval

A prioritized list of unanswered reviews with themed reply drafts, published only after you approve each one.

Pull the Google reviews for [location name] from the last 60 days. Focus on the ones that are still unanswered.

Give each review one theme tag: shipping and delivery, returns and refunds, product quality, sizing or fit, in-store experience, or other. Sort the list in this order: 1-2 stars first, then 3 stars, then 4-5 stars. Within each star group, put the oldest review first.

Draft a reply for every unanswered review using these rules:
- Keep it under 80 words. Write in the reviewer's language. Address the specific point they raised.
- For negative reviews, acknowledge the problem without arguing. Do not promise discounts, refunds or compensation. Direct the reviewer to [customer service contact].
- For positive reviews, thank them and mention the product or service they named. Do not use generic copy-paste lines.

Deliver a table with these columns: date, stars, theme, review excerpt, draft reply. After the table, give a short summary: which theme appears most often in the 1-3 star reviews, and whether that problem should be fixed on product pages or in ad copy before we spend more on ads.

Do not publish anything. I will mark which drafts to edit or approve. Post only the approved replies, and only after I confirm.

SEO and content

Build local service keyword clusters from seed terms

A clustered keyword map for your service and city, ranked by search volume and buying intent, plus a list of negative keyword candidates.

Use Keyword Planner to research keywords for my [service] business in [city]. Seed keywords: [3-5 seed keywords]. Pull average monthly searches and competition for every idea, then:

1. Drop ideas under 20 monthly searches unless they carry a clear buying signal (emergency, near me, price, cost, same day, repair, install).
2. Move DIY, jobs, salary, training, course and "free" terms to a separate list of negative keyword candidates.
3. Group the rest into intent clusters: emergency/urgent, price/cost, service + neighborhood or district, best/comparison, informational.

For each cluster, give me a table with these columns: keyword, monthly searches, competition, suggested landing page type (service page, location page or blog post).

Finish with the top 5 clusters ranked by total monthly searches weighted by buying intent. Count emergency and price terms as high intent and informational terms as low. Add one sentence per cluster on why it matters for calls and leads. This is research only. Do not create or change anything in the account.

SEO and content

Indexing audit: published pages Google is not showing

Finds published WordPress pages with zero Search Console impressions, inspects their index status in batches and groups them by cause with a fix for each, read-only.

List every published page and post on the WordPress site [site name], with its URL and publish date (page through the results until you have them all). Drop anything published in the last 14 days.

Confirm that [property URL] is one of my verified Search Console properties, then pull its page report for the last 90 days with a limit high enough to cover the whole site. Match URLs ignoring http/https, www and trailing slash differences. Flag every remaining published URL that has zero impressions or does not appear in the report at all.

Inspect the flagged URLs in batches of 20, newest first, up to [maximum URLs to inspect, e.g. 100] in total. Stop early and tell me if you hit a quota error. For each inspected URL, read the canonical we declare from its WordPress SEO fields. Return a table with these columns:
- URL
- publish date
- verdict and coverage state
- last crawl date
- robots/indexing state
- canonical we declared
- Google-selected canonical

Group the inspected results:
1) Crawled but not indexed
2) Discovered but never crawled
3) Excluded by noindex or robots
4) Canonical points elsewhere (Google picked a different URL than we declared)
5) Indexed but getting no impressions

For each group, give the likely cause and the fix. Fixes include merging thin content, removing a wrong noindex, adding internal links from pages that already get traffic, or redirecting a duplicate. For the 3 most important URLs in groups 1 and 4, run a full single URL inspection for the extra detail.

Report the counts so I can show management the scale of the problem: total published pages older than 14 days, how many have zero impressions, how many you inspected, and how many of the inspected ones fall into each group. Do not extrapolate to the uninspected URLs. List them separately as "not inspected yet" so I can run the next batch later.

This is read-only. Make no changes.

SEO and content

Diagnose an organic traffic drop: rankings, demand or CTR

Compares Search Console before and after a drop date, classifies each losing page as a ranking, demand or CTR loss, checks the top suspects and delivers a one-page memo with the first fixes.

Organic clicks on [property URL] dropped around [date of drop]. Compare the 14 days after that date with the 14 days before it in Search Console. Search Console data lags about 2 to 3 days, so if fewer than 14 complete days have passed since the drop, use the days available on both sides and tell me.

1. Confirm the property, then pull the daily click series for the period so we can see the exact day the drop started. If it clearly started on a different day, use that day and say so.
2. Check which breakdowns the Search Console report supports, then pull both 14-day windows by explicit start and end dates, broken down by page, by query, by device and by country (one call per breakdown and window). Show the top movers for each.
3. For the 20 pages that lost the most clicks, classify each loss with this rule:
- Ranking loss: average position worsened by 2 or more and impressions fell.
- Demand loss: position is stable (within 1) and impressions fell by more than 20%.
- CTR loss: position and impressions are stable but CTR fell by more than 25%.
- Mixed: anything that fits none or more than one of these.

For the 5 biggest ranking losses:
- inspect the URL for index, canonical and crawl problems,
- read the stored rank history (it is free) for those pages' main queries if they are tracked, and say which ones are not.

For demand losses, check seasonality on up to 5 of the affected main queries with keyword trends before blaming the site. This costs SEO add-on credits, so first check whether the same query was already run and can be reused for free, and skip it if the add-on is not available.

For the top 3 CTR losses, optionally check whether Google currently shows an AI Overview for the main query and whether my site is cited (add-on credits). This shows today's results page only, not what it looked like before the drop, so treat it as supporting evidence.

Deliver a memo of at most one page:
- the dominant cause with the numbers behind it,
- a table of the affected pages with clicks, impressions, CTR and position for both windows and their classification,
- the 3 fixes to do first.

Mark every statement as either measured (it comes straight from the data) or hypothesis. Attributing a CTR loss to a SERP feature or AI Overview is always a hypothesis.

SEO and content

Organic landing pages ranked by business value, Search Console plus GA4

Joins Search Console page data with GA4 organic landing page sessions and key events, sorts pages into push, fix, protect and deprioritize groups, and names the 3 pages with the biggest upside.

Management wants to know which organic pages actually bring business, not just traffic.

1) List the GA4 key events for [GA4 property] and confirm which ones count as business conversions: [key event names]. Tell me if the property has other key events too, because GA4 reports key events as one combined total.
2) For the last 90 days, pull Search Console clicks, impressions and average position per page for [property URL], with a limit high enough to cover all pages with clicks.
3) For the same 90 days, pull GA4 sessions and key events by landing page split by channel, and keep only the Organic Search rows.
   - If the business events are the only key events in the property, use that key event count directly.
   - If there are other key events, also pull landing page by event name and use it to estimate the share of each page's key events that are business events. Say clearly that this split covers all channels, so the organic figure is an estimate.
4) Join the two sources by URL path: strip the hostname, query strings and fragments, lowercase, and treat a trailing slash as the same page. List the pages that appear in only one source. Build a table with these columns:
   - page
   - organic clicks
   - average position
   - organic sessions
   - key events (business only)
   - conversion rate (key events divided by organic sessions)

Leave pages with fewer than 30 organic sessions out of the median calculation, since their conversion rate is noise, and list them separately.

Then sort the pages into four groups. If I have not given a [click threshold], use the median organic clicks of the included pages and tell me the number.
- Push SEO: conversion rate above the site median but fewer clicks than the threshold.
- Fix conversion: clicks above the threshold but conversion rate below half of the median.
- Protect: clicks above the threshold and conversion rate above the median.
- Deprioritize: everything else.

Add a short note under the table on data limits: Search Console clicks and GA4 sessions are counted differently, consent banners hide part of GA4 traffic, and GA4 attributes key events to the landing page of the session.

Close with the 3 pages where one improvement (better rankings for Push SEO pages, a better offer or form for Fix conversion pages) would add the most conversions, with a rough estimate of the gain and how you calculated it.

SEO and content

Cluster my Search Console queries and find keyword cannibalization

A table of topic clusters built from your real organic queries, with overlapping pages flagged and the easiest page-2 wins listed.

SEO and content

Find the keyword clusters my competitors rank for and I don't

The top 10 missing topic clusters versus two competitors, with the competitor page that wins each one and what you should build.

SEO and content

Map keyword clusters to my WordPress pages and fix SEO titles

A one-cluster-per-page map of your WordPress site with conflicts flagged and new meta titles and descriptions applied only after approval.

For my WordPress site [site URL], list the published pages and posts. Read each one's meta title, meta description and focus keyword.

Pull Search Console queries per page for the last 90 days. Group each page's queries into its main topic cluster.

Build a table with one row per page:
- URL, current focus keyword and current meta title
- main cluster it actually ranks for, with impressions and average position
- status:
  - OK: the focus keyword matches the main cluster
  - Mismatch: the page ranks for a different cluster than it targets
  - Missing: no focus keyword or meta title is set
  - Conflict: another page targets the same cluster

For the 10 pages with the most impressions that are not OK, draft:
- a new meta title of up to 60 characters, with the main cluster keyword near the start
- a meta description of up to 155 characters, with a clear reason to click

Show all drafts as a before/after preview and wait for my approval. After I approve, apply only the approved changes and purge the cache. For Conflict pages, give a recommendation only and make no change.

SEO and content

Ranking losses on pages that actually sell, with root cause

Finds the organic ranking and click losses of the last 28 days, weights them by GA4 organic purchases per landing page, and diagnoses the top pages with URL inspection and a live SERP check. It uses stored rank history when it exists and falls back to Search Console when it does not. Read-only.

Find which ranking losses are costing us sales on [domain] (Search Console property: [site], GA4 property: [GA4 property]).

Step 1: find the losses.
- Read the stored rank history for [domain] over the last 30 days (free). List keywords that dropped out of the top 10, or lost 3 or more positions.
- If there is no stored history, or it covers only a few keywords, say so plainly and use Step 2 as the main source of losses. Get current positions from the domain's ranked keywords (check seo_history first so we reuse a recent run).

Step 2: confirm with Search Console.
- Pull page,query rows for the last 28 days and for the last 56 days. Derive the previous 28 days by subtraction: clicks and impressions = 56-day value minus 28-day value. Previous position = (56-day position × 56-day impressions − 28-day position × 28-day impressions) / previous impressions.
- Flag page,query pairs that lost 3 or more positions, or fell from top 10 to beyond 10, and calculate click change per page.

Step 3: weight by sales.
- List the key events and confirm the purchase event name.
- Pull GA4 for the last 90 days by landing page with channel as the second dimension, and keep Organic Search for sessions and key events. If the property has several key events, also run landing page × event so the purchase count is isolated. If organic and purchase cannot be isolated together, say which number you used.

Deliver one table, sorted by organic purchases, with these columns:
- landing page
- lost keywords
- average position change
- click change %
- organic purchases (90 days)
- priority

Priority rule:
- High: the page has ≥ [minimum purchases] organic purchases and lost ≥ 20% of clicks.
- Medium: one of those two conditions is true.
- Low: the rest.

For the top 5 high-priority pages:
- Inspect the URL in Search Console: indexing, chosen canonical, last crawl.
- Run a live SERP check on the biggest lost keyword to see who replaced us and with what kind of page.

Give a one-line likely cause and next action per page. If the stored history was missing, end with a list of the 20 keywords I should start tracking, but do not start tracking them. This is read-only, so change nothing.

SEO and content

Spam and PBN link review with a draft disavow list

A keep / watch / disavow-candidate table of risky links, checked against organic traffic trends, plus a draft disavow file for manual upload.

Review the backlinks of [my domain] for spam and link-network patterns.

1. Pull the individual backlink list and run a bulk spam score on all referring domains.
2. Check referring IPs and subnets. Flag any subnet hosting 3 or more of my referring domains.
3. Flag links that match any of these: spam score above 60, foreign-language sites unrelated to my service, sitewide footer or sidebar links, exact-match commercial anchors on low-quality sites.
4. Pull Search Console organic clicks by week for the last 6 months. Tell me whether any drop lines up with a spike in flagged domains.

Deliver a table with columns: domain, spam score, subnet, flag reason, verdict (keep / watch / disavow candidate).

Be conservative. Most low-quality links are simply ignored by Google, so mark a domain as a disavow candidate only when there is a clear pattern or a traffic drop that lines up with it.

Then draft the disavow file text (domain: lines) for the candidates only. I will review it and upload it myself in Google's disavow tool.

SEO and content

Find top queries where an AI Overview appears but my site is not cited

A table of your top organic queries showing AI Overview presence, whether you are cited, who is cited instead, and the at-risk queries to fix first.

SEO and content

How much traffic and how many leads do AI assistants send me?

A GA4 breakdown of sessions and key events from ChatGPT, Perplexity, Gemini, Copilot and Claude, compared with Google organic, plus the landing pages that matter.

SEO and content

Add structured data to my hosted site so AI engines understand it

A JSON-LD gap check for your homepage and key service page, with exact code edits shown as a diff and applied only after approval.

SEO and content

Brief a new answer-style page for questions I almost rank for

A content brief for one new page targeting question queries stuck on page 2, built to be cited in AI answers and saved as a WordPress draft only on approval.

From Search Console, find the question-style queries for [site URL] over the last 90 days. These are queries that contain how, what, why, which, cost, price or "vs". Keep only those with an average position between 8 and 30 and at least 100 impressions.

Group them by topic and pick the single topic with the most impressions. Before going further, check my existing WordPress pages. If I already have a page on this topic, tell me, and suggest improving that page instead of creating a new one.

For the chosen topic:
1. Check whether an AI Overview appears and which domains AI answers cite.
2. Generate a content brief.

Deliver a brief that includes:
- the target query and the secondary queries
- the H1
- a 40 to 60 word direct answer for the top of the page
- an H2 outline phrased as questions
- the facts, numbers or table that the cited pages include
- a 5 question FAQ
- 3 internal links to existing pages on my site
- the recommended length

After I approve the brief, create the page in WordPress as a draft only. Do not publish it.

SEO and content

Expand a ranking page with the queries it already gets but does not cover

A query-by-query coverage table for one URL plus draft sections for the biggest gaps, ready to approve.

Work on the Search Console property [property] and the page [URL] on the client's WordPress site.

1. Pull the last 90 days of Search Console data for this page by query. Keep queries with at least 100 impressions and an average position between 5 and 20.
2. Read the full page content from WordPress and check whether each query, or a close variant, is answered in the headings or body.
3. Deliver a table with these columns: query, impressions, clicks, CTR, average position, coverage (yes / partial / no), and recommended action. The action must be one of: new H2 section, expand existing section, add FAQ item, or ignore as off-topic.
4. Draft new content for the 5 uncovered or partially covered queries with the most impressions. Write an H2 and 80 to 150 words for each, in the page's existing tone. Use concrete facts only and no filler.
5. Show where each block would be inserted in the current page.

Do not edit the page yet. Present the updated content as a preview and wait for my explicit approval before updating the post.

SEO and content

Find and resolve keyword cannibalization between client pages

Uses Search Console page and query data to find URLs splitting the same query, decides merge, differentiate or canonical per case, and applies redirects and canonicals only after preview and approval.

Check the Search Console property [property] for keyword cannibalization over the last 90 days.

1. Confirm the property with gsc_list_sites. Run list_servers to see whether the site is on Opus hosting or a connected WordPress site.
2. Pull gsc_performance with dimension page,query for 90 days, sorted by impressions, with a high limit (page through if needed). From these rows, calculate each URL's share of impressions per query. Keep queries with at least 200 total impressions where two or more URLs each get at least 10%.
3. For each case, show every competing URL with its impressions, clicks, CTR and average position for that query, plus how many other queries each URL earns 50 or more clicks on.
4. Read the competing pages (wordpress_get_post on WordPress, site_read_url with outline=true on hosting) and the current SEO fields with wordpress_seo. Summarise in one line what each page targets.
5. Decide per case:
   - If one URL has 70% or more of the clicks and the other has no unique query with 50 or more clicks, merge the weaker page's useful sections into the stronger one and 301 redirect the weaker URL.
   - If the pages serve clearly different intents, keep both and specify the title, H1 and section changes that separate them.
   - If they are near-duplicates that must both stay live, set a canonical on the weaker page to the stronger URL.
6. Deliver a table with these columns: query, URLs, metrics, decision, exact changes. Order by clicks at stake.

Do not apply anything yet. Show the full list of content merges, redirects and canonical changes as a preview and wait for my approval. Then:
- Opus hosting: preview redirects with site_redirects (mode add, confirm=false), apply with confirm=true. Change canonical tags with site_edit_file after reading the file with site_read_file.
- WordPress: set canonicals with wordpress_set_seo. Apply approved content merges with wordpress_update_post. For redirects, check wordpress_list_plugins: if a redirect plugin is active, give me the source and target list to enter there. If not, offer a WPCode snippet that issues the 301s on template_redirect via wordpress_snippet, shown as a preview with confirm=false before creating it. Run wordpress_purge_cache afterwards.
Finally verify each redirected URL returns 301 to the right target with site_read_url.

SEO and content

Local service page briefs from Business Profile search terms

A map of how locals search for the business against existing service pages, plus briefs for the missing ones.

For the Business Profile location [location] and the Search Console property [property], check whether the website has a page for every service people actually search for.

1. Pull the search terms that led to the Business Profile over the last 3 months. Read the listing's categories and services.
2. Group the terms into service clusters. Leave out pure brand searches.
3. List the published pages on the client's WordPress site. Match each cluster to an existing service page, or mark it as missing.
4. For matched pages, add the page's organic clicks and average position for that cluster from Search Console over the last 90 days.
5. Deliver a table with these columns: service cluster, example terms, search count, matching URL, organic position, and action. The action must be one of: create page, strengthen page, or add to the Business Profile services list.
6. For the top 3 missing pages by search count, write a brief. Each brief needs:
   - a title tag (max 60 characters) and an H1 with the service and city,
   - an H2 outline,
   - a local proof section using 2 or 3 real review excerpts from the listing,
   - an FAQ with 4 questions,
   - LocalBusiness and Service schema notes.

This is analysis only. Do not edit the listing or the site.

SEO and content

Align GBP categories and services with the product searches that matter

A gap analysis between your listing's categories and services and your highest-impression organic product queries, with changes to approve.

Check whether the categories and services on [location name] reflect what people search for before buying from us.

1. Read the listing's primary category, additional categories and service list.
2. Pull Business Profile search keywords for the last 3 months. Also pull the Search Console queries for [Search Console property] for the last 90 days. Keep only queries with more than 100 impressions that name a product type or a local modifier, such as a city, a district or "near me".
3. Group those queries into product themes. For each theme, show total impressions, clicks and average position. Then mark whether the listing covers the theme with a category, a service, both or neither.

Decision rules:
- If a theme has high impressions and no matching category, search for the closest Business Profile category and suggest it as an additional category. Never suggest changing the primary category without explaining the risk.
- If a theme has no matching service, draft a service name and a description of up to 300 characters.

Show the proposed category and service changes as a preview. Wait for my approval before updating the service list. List any category changes separately as manual steps for me.

SEO and content

Assign keyword clusters to WordPress pages and fix meta titles

Maps Search Console query clusters to your WordPress service and location pages, flags mismatches and missing pages, and rewrites weak meta titles and descriptions after your approval.

Map my keyword clusters to pages on my WordPress site [site]. My services are [service list] and the cities or districts I serve are [locations].

1. Confirm the site is connected, then list all published pages and posts (page through the results until you have them all). Classify a page as a service page if its title or slug names one of my services, and as a location page if it also names one of my locations. Put everything else in "other" and leave it out of the work below.
2. For every service and location page, read its current meta title, meta description and focus keyword.
3. Find my Search Console property for [site] and pull the last 90 days of performance by page and query together, sorted by impressions, with a limit high enough to cover all my pages.
4. Match the Search Console page URLs to the WordPress pages by their link or slug, ignoring http/https, www and trailing slash differences. List any URL you could not match.
5. Group each page's queries into clusters of the same intent (service plus location, service only, informational). Give each page one primary cluster: the cluster with the most impressions that fits the page's content. Name the primary keyword of that cluster.

Flag three problems:
- pages with no clear cluster (fewer than 100 impressions in total)
- pages whose focus keyword differs from the cluster they actually rank for
- clusters with at least 200 impressions that have no dedicated page (the queries spread across several pages or land on the home page)

Draft new SEO fields only where they are needed:
- If the current meta title already contains the primary keyword and is 60 characters or less, mark the page "OK" and leave it alone.
- Otherwise, meta title: maximum 60 characters, containing the service and the city.
- Meta description: maximum 155 characters, ending with a clear call to action such as calling now or requesting a free quote. Only rewrite it if it is missing, over 155 characters or does not mention the service.
- Where the focus keyword is wrong, propose the primary keyword as the new focus keyword.

Show all proposed changes as a before and after table with page, type, primary cluster, impressions, current and new title, current and new description, current and new focus keyword, and character counts. Write nothing until I approve. After my approval, apply only the rows I approved (overwrite existing values only for those), read the SEO fields back to confirm they landed, then purge the page cache.

SEO and content

Fix low-CTR pages that already rank on page one

A list of pages that rank well but underperform on CTR, plus rewritten meta titles and descriptions applied in WordPress only after approval.

Pull page and query data from Search Console for [property URL] for the last 90 days.

1) Find the site's own benchmark. Group queries by average position bucket (1-2, 3-4, 5-6) and calculate the median CTR in each bucket.
2) Flag the low-CTR pages. A page qualifies when all three are true:
   - it has at least 1,000 impressions,
   - its average position is 6 or better,
   - its CTR is below 60% of the median for its bucket.
   Rank the flagged pages by missed clicks, calculated as impressions × (bucket median CTR - actual CTR).
3) Read the current SEO fields of the top 10 pages on the WordPress site [site name] (meta title and meta description) and their top 3 queries.
4) Propose a new meta title of at most 60 characters and a new meta description of at most 155 characters for each page. Each one should match the search intent of the main query. Use no clickbait and make no claims the page does not support.

Show a before/after table with URL, current fields, proposed fields and missed clicks. Do not change anything yet. Wait for me to approve each row. After that, apply only the approved rows and purge the cache.

SEO and content

Page-two quick wins: ranked plan to move queries into the top 5

A prioritized list of queries ranking 8-20, with click potential and a concrete content action for each target page.

Using Search Console for [property URL], last 90 days, find non-branded queries with an average position between 8 and 20 and at least [minimum impressions] impressions.

For each query, estimate the extra clicks it would gain if it moved to position 3. Use the site's own average CTR at positions 1-3 as the benchmark instead of a generic industry curve.

Keep only queries where a single page receives most of the impressions. If two or more pages compete for the same query, list them separately as cannibalization cases.

Take the top 10 opportunities and, for each one:
- run an on-page audit of the ranking URL,
- generate a content brief for the main query,
- name the single most important action: expand a section, add an FAQ block, improve the title, add internal links from named pages, or fix a technical issue.

Deliver a ranked table with query, URL, impressions, current position, estimated extra clicks, action and effort (low, medium or high).

This is a plan only. Do not edit any page.