If AI answer engines are recommending your pages but those pages aren’t earning clicks in Google, you have a clear content and SEO opportunity — and a blind spot. This dashboard pairs prompt-level AI citations (from Semrush) with page-level Google performance (from Search Console) so you can find pages AI trusts that Google users ignore, prioritize fixes, and track progress each week.
What this dashboard shows and why it matters
The dashboard brings two viewpoints together: where your brand appears inside AI answers, and how the same landing pages perform in Google Search. That lets you answer practical questions fast: which pages do AI systems cite for buyer queries, are those pages converting organic interest into clicks, and where should editorial effort go first?
Key metrics the dashboard combines:
- AI visibility rate — share of tracked prompts where your brand appears (Semrush Prompt Tracking)
- Prompt-level mentions and the specific pages AI cites (Semrush Visibility Overview)
- Generative AI impressions for your pages (Google Search Console’s Generative AI report, when available)
- Page-level clicks, impressions, CTR and average position from Google Search Console (standard Performance report when Generative AI data isn’t present)
Example from testing: a guide that AI answers cited for 50 tracked prompts but that earned zero Google clicks — a high-priority item to rework for intent and click potential.
What you need before you start
- Semrush Prompt Tracking / AI Visibility Toolkit (subscription for prompt tracking and visibility exports).
- Google Search Console property access to export page-level performance or Generative AI report data.
- A Google Sheet to consolidate exports and keep a stable data structure.
- A paid Claude plan (Pro, Max, Team or Enterprise) that supports the Google Drive connector and artifacts.
- Editor time: the workflow typically takes ~90 minutes to build initially and ~20 minutes for weekly updates.
Step-by-step: assemble a working dashboard
1) Define what you will measure
Decide the target domain and brand variants, which AI platforms to track (Semrush supports ChatGPT, Google AI Mode and Gemini in this workflow), the languages and locations, and a stable set of prompts that reflect how buyers actually ask questions. Choose a reporting window (for example, rolling 28-day comparisons). Crucially: keep the prompt set stable for at least a month — swapping prompts breaks trend continuity.
2) Build a representative prompt set
Create a prompt-tracking campaign in Semrush and add prompts that mirror real customer queries across the funnel: problems, discovery, comparisons, product selection, and branded questions. Start with ~30 prompts across five groups (roughly six per group) to get enough observations for a stable site-level visibility rate. Tag prompts by group so you can segment trends rather than reporting one blended score.
3) Export Semrush AI datasets into Google Sheets
Export two datasets into your master sheet: prompt-level results (one row per prompt per day with position, answer type and cited landing page) and the Visibility Overview exports for “Cited Sources” and “Cited Pages.” These specific exports are currently available only via the Semrush UI, so use scheduled exports or manual downloads to populate your Sheet.
4) Export Google Search Console data
If your property has data in GSC’s Generative AI report, export daily impressions (three-month window) for trends and the pages table for a 28-day window. If that report is empty or thin, export page-level clicks, impressions, CTR and average position from the standard Search results performance report for matched 28-day windows. End export windows 2–3 days before today to avoid GSC’s freshest incomplete rows.
5) Consolidate everything in a single master Google Sheet
Keep a fixed tab structure and consistent headers so downstream automation can parse reliably. Suggested tabs: semrush_prompts (one row per prompt per day), semrush_cited_pages, semrush_source_domains (optional), gsc_genai_daily, gsc_genai_pages, and gsc_pages for standard exports. Record a snapshot date for Semrush exports — Visibility Overview can include historical mentions even for a newly created campaign.
6) Turn the Sheet into a live artifact in Claude
Connect Google Drive to Claude via its connector, confirm Claude can open the master Sheet, and ask Claude to create an artifact that reads the Sheet live. The artifact will visualize prompt visibility trends, cited pages and GSC metrics side-by-side. Note sharing limits: artifacts that call connectors require a paid Claude plan and cannot be shared publicly; Team and Enterprise accounts can share internally subject to viewer access to the Sheet.
7) Test and validate update mechanics
Edit a known value in the Sheet (for example clicks in the gsc_pages tab), refresh the artifact and confirm the dashboard reflects the change and shows a data-freshness stamp. If panels go blank after edits, check that column headers and tab structure exactly match the template.
How to use the dashboard: priorities and signals
Use the combined view to triage content work quickly:
- Pages AI cites but that earn few Google clicks: low-effort, high-impact fixes. Improve organic relevance, target a realistic keyword, refresh headings and meta tags, and prioritize link-building where appropriate.
- Prompts dominated by competitor mentions: analyze competitor answers and create targeted comparison or category content that directly addresses the prompt’s intent.
- Prompt groups with no presence: these indicate missing top-of-funnel coverage — produce brief content that answers those specific user needs.
Also watch for: pages that rank well in Google but aren’t cited by AI — make answers more extractable (concise headings and direct answer passages). If a page loses AI mentions after edits, review the changes for removed answerable content. Positive movement in both AI mentions and Google clicks is the strongest signal that your work is aligning AI visibility with organic demand.
Limitations and operational notes
The workflow depends on vendor exports and a paid artifact connector. Semrush prompt and cited-page exports are not currently available via API, limiting full automation to scheduled exports or manual steps. Artifacts that call Drive connectors require a paid Claude plan and cannot be shared by public link — for stakeholder distribution, export the artifact’s report as HTML.
Also, Google’s Generative AI metrics in Search Console do not include query-level data or prompt-to-click attribution, so Semrush provides prompt-level visibility while GSC provides click behavior; you will not get direct prompt-to-click mapping from Google today.
Next step: lock your prompt set, run the first two weekly exports to confirm the data pipeline, and prioritize pages AI cites but users aren’t clicking. Those pages are often the fastest route to measurable gains in both AI visibility and organic traffic.
