Building an AI Visibility Dashboard with Semrush and Claude

Semrush users can build a custom AI visibility dashboard to track how often AI platforms cite their brand relative to organic search performance. This dashboard compares prompt-level data from the Semrush AI Visibility Toolkit with page-level data from Google Search Console (GSC) to identify content gaps. (AI Visibility Toolkit: a software tool from Semrush used to track how often brand names appear in different artificial intelligence platform responses.) The first build takes 90 minutes, while subsequent weekly updates take 20 minutes.

SEO managers monitoring brand visibility should use this system to correlate AI visibility rates—the share of tracked prompts where a brand appears—with actual Google click volume. By identifying pages that AI platforms trust but users do not click, managers can determine if a content piece requires better Google ranking or a structural AEO (Answer Engine Optimization) update.

Required tools and technical constraints

Building this dashboard requires a specific software stack: Semrush AI Visibility Toolkit, Google Search Console, Google Sheets, and a paid Claude plan. The Claude subscription requirement is necessary because the dashboard functions as a Claude artifact—a live, interactive UI component—that pulls data through the Google Drive connector. While the connector is available to all users, only paid plans (Pro, Max, Team, or Enterprise) support artifacts that call external data sources.

The dashboard relies on a master Google Sheet to hold and structure the data.

The dashboard relies on a master Google Sheet to hold and structure the data. Because Semrush does not currently offer an API for these specific datasets, the workflow depends on manual exports. Users must ensure their Google account can open the master Sheet for the Claude artifact to function correctly.

Mapping the buyer journey with 30 strategic prompts

Data accuracy depends on a stable set of prompts that mirror how actual buyers talk to AI. Semrush recommends starting with 30 prompts distributed across five distinct groups to avoid skewed results. These groups include customer problems, product or category discovery, comparisons, product selection, and branded questions. Spreading prompts across the full journey provides a more stable site-wide visibility rate than clustering on brand terms.

The non-branded prompt results

Each group should eventually contain at least 15 prompts to stabilize reporting, as a single prompt change in a group of six moves that group’s rate by 17 points. The non-branded prompt results are particularly important because they reveal whether AI recommends the brand to users who do not yet know it. Users must keep this core prompt set stable for at least one month to ensure trend lines remain valid.

The data export and normalization workflow

Extracting data requires two separate exports from the Semrush AI Visibility Toolkit: one for prompt-level performance and one for cited pages and sources. For the Google Search Console data, users must check if the Generative AI performance report contains data for their property. (Generative AI performance report: a specific report within google search console that shows how content performs within artificial intelligence search results.) If the report is empty, users should use the standard “Search results” performance report instead.

Google Search Console data

Google Search Console data should be exported for 28-day windows, ending 2-3 days before the current date to ensure data completeness. All URLs must be normalized across the Google Sheet tabs to prevent GSC from reporting the same page multiple times due to trailing slash variations. If page-level numbers appear unnaturally low, the first check should be for inconsistent URL formatting.

Building the live dashboard with Claude artifacts

Claude artifacts turn the static Google Sheet into a live reporting layer. To set this up, connect the Google Drive to Claude and provide a specific prompt that instructs Claude to build the dashboard. This prompt should define the connection to the specific sheet name and the domain being tracked.

The artifact includes a data-freshness stamp to prevent users from viewing stale information.

The artifact includes a data-freshness stamp to prevent users from viewing stale information. Because the artifact reads the live sheet on demand, the dashboard updates automatically whenever the Google Sheet is modified. If a section of the dashboard goes empty after an update, the user must compare the column headers in the Sheet against the original structure to ensure they match exactly.

Prioritizing actions based on dashboard findings

The dashboard identifies three specific types of content opportunities based on the correlation between AI citations and organic clicks. A page cited 50 times by AI but earning zero Google clicks indicates a trust win that needs a Google-specific ranking refresh. Conversely, a page that ranks well in Google but is rarely cited by AI may need a structural overhaul, including more direct answers and descriptive headings.

The dashboard also identifies competitors cited for prompts where the brand

The dashboard also identifies competitors cited for prompts where the brand has no presence. These prompts provide a roadmap for new comparison and category content. Finally, if an entire prompt group shows zero visibility, it signals a total content gap that requires new top-of-funnel pieces based on those specific customer questions.

Maintaining the dashboard with a weekly update routine

1. Export the latest prompt tracking data from the Semrush campaign.
2. Export the latest Cited Pages and Cited Sources tables from Visibility Overview.
3. Export the latest GSC data, helping ensure the window ends 2-3 days prior to on 12 October 2026.
4. Append the new rows to the master Google Sheet with a fresh snapshot date.
5. Refresh the Claude artifact to update the visuals and data-freshness stamp.

Who this is for: SEO managers tracking brand visibility

SEO managers tracking brand visibility should use this to identify which pages need a structural AEO update versus a traditional backlink and ranking push. Growth engineers looking to automate reporting can use the Claude Drive connector to create a self-updating dashboard that eliminates manual chart building. Content strategists can use the identified citation gaps to prioritize which pages to update or which new topics to cover in their content plans.

Who this is not for

Beginner bloggers should skip this method because it requires a paid Claude plan and the Semrush AI Visibility Toolkit. Social media managers can ignore this as the dashboard focuses exclusively on organic search and generative AI engines rather than social platforms; these users should stick to standard social media analytics.

Leila’s estate planning scenario

Leila owns an estate planning firm and wants to see how AI handles queries about settling a father’s estate. She sets up a campaign with 30 prompts, including six for customer problems like “I was just named executor of my dad’s estate.” After the first 90-minute build, the dashboard reveals that her “executor checklist” is cited for 50 prompts but earns zero Google clicks. Leila prioritizes a Google-focused ranking refresh for that page to capture the traffic the AI is already sending to her competitors.

If Leila had found zero visibility for those same six customer-problem prompts, she would have pivoted instead to creating entirely new top-of-funnel content to address those specific questions. The dashboard’s ability to distinguish between a ranking problem and a trust problem changes her entire content priority list.

What this can be used for

What this replaces

This workflow replaces the need for manual data merging in Excel and the creation of static PDF reports. Because the Claude artifact connects directly to the Google Drive, it automates the visualization process and helps ensure the dashboard is live rather than a point-in-time snapshot. This reduces the time spent on manual chart construction and helps ensure stakeholders always see the most recent data.

What is not yet known: Semrush does not

Semrush does not state which specific regions, languages, or devices are most impacted by these visibility shifts. It is also unknown if site owners can opt out of being cited by these AI platforms or what specific brand-name variations are most common for every brand. These variables would need to be tested through additional regionalized campaigns or brand-specific monitoring.

Reported by Semrush on 12 October 2026: How to build an AI visibility dashboard with Semrush.