An AI visibility dashboard exhibits you whether or not AI platforms cite and advocate your model.
We’ll assist you construct one which places your prompt-level AI knowledge from Semrush subsequent to your page-level Google knowledge from Search Console, so you’ll be able to see whether or not the pages AI cites are additionally incomes clicks.
On our check website, the primary refresh surfaced a information that AI solutions cited for 50 prompts and that earned zero Google clicks.
To construct your individual model visibility dashboard for AI search, you’ll want Semrush’s AI Visibility Toolkit, Google Search Console (GSC), Google Sheets, and a paid Claude plan. The Sheet holds your knowledge. Claude turns it right into a stay, interactive dashboard. It took me about 90 minutes to construct out this dashboard the primary time and it takes 20 minutes to replace every week with recent knowledge.
What this AI visibility dashboard will present
This dashboard exhibits two measurements aspect by aspect: how seen your model is in AI solutions (from Semrush) and the way your pages carry out in Google (from GSC). Particularly, it tracks:
- AI visibility charge: the share of your tracked prompts the place your model seems
- Mentions and cited pages from Semrush’s immediate monitoring
- Generative AI impressions from GSC’s Generative AI efficiency report
- Clicks, impressions, click-through charge (CTR), and common place from GSC
- An inventory of website positioning and reply engine optimization (AEO) alternatives based mostly in your actual knowledge
There are two constraints to concentrate on earlier than you start:
- The dashboard runs as a Claude artifact that pulls stay knowledge out of your Google Sheet by way of the Google Drive connector. The connector itself is on the market to everybody, however artifacts that decision connectors require a paid Claude plan (Professional, Max, Group, or Enterprise).
- You’ll be able to’t share it by public hyperlink. On Group and Enterprise plans, you’ll be able to share it inside your group, and every viewer sees knowledge provided that their Google account can open your Sheet and their Claude account has the Drive connector. On Professional and Max, it stays non-public to you.
This can be a working dashboard for whoever owns reporting, not a hyperlink you ship to management. For stakeholder reporting, you’ll be able to ask Claude to show the dashboard’s report into an HTML file they’ll open in any browser even with out the connections.
Step 1: Determine what you wish to measure
Earlier than you open any device, resolve what you wish to measure:
- Your goal area, plus brand-name variations AI would possibly use
- Which AI platforms you wish to monitor (Semrush’s Immediate Monitoring helps ChatGPT, Google AI Mode, and Gemini)
- Location and language
- A steady set of prompts tied to how consumers really discuss to AI
- Your required reporting interval, akin to the newest 28 days versus the earlier 28 days
Step 2: Arrange AI immediate monitoring in Semrush
To arrange AI immediate monitoring, create a marketing campaign within the Immediate Monitoring device, which checks a set set of prompts in opposition to the chosen AI platform each day and information whether or not your model seems.
Select and add your prompts
Add prompts that mirror how your consumers discuss to AI. In the event you can, mine buyer help tickets, gross sales calls, and discussion board posts. You’ll be able to feed transcripts and textual content posts into Claude to get higher immediate choices based mostly on the patterns it sees.
Unfold your prompts throughout the complete journey fairly than clustering on model phrases. I used 30 prompts, six in every of 5 teams:
- Buyer issues: The problems your potential prospects are coping with (e.g., “I used to be simply named executor of my dad’s property and do not know the place to begin. What do I do first?”)
- Product or class discovery: When your prospects don’t know your services or products exists but, they usually’re looking for an answer; e.g., “Is there software program that helps executors settle an property?”
- Comparisons: Evaluating your model to a competitor’s model; e.g., “Which is best for [your ideal customer or one specific problem you help solve], [your brand] or [competitor]?”
- Product choice: Researching services or products choices past direct comparability; e.g., “What are one of the best instruments for an property with no will?”
- Branded questions: Questions instantly about your model; e.g., “Is [brand] legit?”
Begin with round 30 prompts. Monitored every day, that is roughly 200 observations per week, sufficient for a steady site-wide visibility charge. The entry AI Visibility Toolkit plan covers 25 prompts, so 30 means a better tier or a immediate add-on.
Group-level numbers will swing. With six prompts per group, one immediate altering strikes that group’s charge by 17 factors, so deal with group adjustments as results in examine. If stakeholders need reliable group-level reporting, broaden every group to fifteen or extra prompts over time, and solely add prompts that mirror actual purchaser language.
Your branded prompts will nearly at all times present your model, which lifts your total charge. Report on all prompts and non-branded solely. The non-branded quantity tells you whether or not AI recommends you to individuals who do not know you but.
Hold the identical core prompts steady for no less than a month as soon as monitoring begins. Each pattern line in your dashboard relies on the immediate set staying fixed. In the event you swap prompts mid-stream, you break the reporting.
Manage your prompts with tags
Tag every immediate with the group it belongs to, so you’ll be able to evaluate visibility throughout significant segments as a substitute of 1 blended rating.
In your marketing campaign, open the “Tags” tab and click on “Handle tags” to create and assign them. I used one tag for every of my 5 immediate teams. The screenshot under is from my Google AI Mode marketing campaign, which tracks a smaller set of prompts throughout the identical 5 teams. Semrush additionally exhibits its personal matter groupings on this view, and you’ll ignore these for the dashboard.
Step 3: Export your Semrush AI visibility knowledge
You’ll must export two datasets from the AI Visibility Toolkit into Google Sheets: prompt-level efficiency from the monitoring marketing campaign you arrange in step 2, and the pages AI platforms cite.
Neither dataset is on the market by way of the Semrush API but, so the information will get to your Sheet by export. You’ll be able to schedule these exports by way of e-mail (extra on that under).
Export prompt-level efficiency
Your immediate monitoring marketing campaign opens in a Place Monitoring view with “AI Visibility” toggled on. As soon as the information is populated (it might take a couple of hours), scroll all the way down to the “Prime Prompts” part and click on “View all X prompts.” Then, click on “Export” within the top-right nook and select Google Sheets.
It’s best to see one row per immediate, with a column group for every day displaying your model’s place within the reply, the reply kind, and the touchdown web page cited. A place of “-” means your model did not seem that day. You will reshape this within the grasp Sheet step.
Export cited pages and sources
In Visibility Overview, use the platform selector to select the identical AI platform your monitoring marketing campaign follows, so your quotation knowledge and immediate knowledge describe the identical platform. Then scroll all the way down to the “Subjects & Sources” part. You’re going to export the Google Sheets knowledge for each the “Cited Sources” and “Cited Pages” tabs.
Visibility Overview holds months of historical past even if you happen to created your immediate monitoring marketing campaign yesterday, as a result of Semrush builds it from its personal ongoing evaluation of the AI reply house. My check website had a mentions pattern going again six months on day one.
What to do if export is unavailable
If the export button is grayed out or lacking, verify your entry to the marketing campaign. If another person owns it, ask them to export the studies for you or to alter your permissions. Nothing else on this workflow wants proprietor rights.
Attain out to Semrush help if you happen to encounter any difficulties right here.
Step 4: Export Google Search Console knowledge
GSC knowledge exhibits what your pages really earn in Google, subsequent to what AI solutions say about you. Which report you utilize relies on whether or not the Generative AI efficiency report has knowledge on your property.
In case your Generative AI report has knowledge
Open your property in GSC, go to “Efficiency,” and open the “Generative AI” report. Export it twice. As soon as with the date vary set to the final 3 months, for the every day pattern. As soon as set to the final 28 days, for the pages desk, so it matches your commonplace GSC window.
The export is a ZIP of separate one-dimension tables (every day impressions, pages, international locations, units). Impressions are presently the one metric accessible.
There isn’t a prompt-to-click attribution; Google presently doesn’t provide you with queries, clicks, or CTR knowledge. The pages desk is aggregated throughout the entire date vary fairly than damaged out by day.
In case your Generative AI report is empty
The Generative AI report is on the market to each property, nevertheless it exhibits knowledge solely as soon as your pages get sufficient impressions in AI Overviews and AI Mode. If yours is empty or skinny, use the usual “Search outcomes” efficiency report as a substitute. Export page-level clicks, impressions, CTR, and common place for 2 matched 28-day home windows.
Finish every export window 2-3 days earlier than at this time, as a result of the most recent GSC knowledge is incomplete and can undercount. Hold the home windows the identical size so interval comparisons are constant.
Observe that that is total Google Search efficiency, not AI-specific visibility, so label it that approach in your dashboard.
Step 5: Manage every thing in a single grasp Google Sheet
Put each export into one Google Sheet with a set tab construction. Mine ended up with seven knowledge tabs, every holding one dataset:
- semrush_prompts: prompt-level outcomes reshaped to 1 row per immediate per day, with columns for date, immediate, platform, tag, place, and cited touchdown web page
- semrush_cited_pages: the cited pages export, stamped with a snapshot date
- semrush_source_domains and semrush_topics: the Visibility Overview exports, identical stamping (subjects is optionally available)
- gsc_genai_daily and gsc_genai_pages: the every day desk out of your 3-month Generative AI export and the pages desk out of your 28-day one
- gsc_pages: commonplace page-level GSC knowledge, one row set per 28-day window
Here’s a immediate you’ll be able to feed to Claude together with all of the exports to have it construct the sheet for you:
On future weekly updates, I am going to connect the present grasp sheet (downloaded as .xlsx) plus new exports.
Step 6: Flip the Google Sheet right into a stay Claude artifact
Subsequent, we’ll flip the Sheet right into a dashboard by asking Claude to construct a stay artifact that reads the Sheet by way of your Google Drive connector.
Join the Google Sheet to Claude
Join Google Drive in Claude’s connector settings, then affirm Claude can open the grasp Sheet earlier than asking for something visible. The connector lets the artifact re-read the stay Sheet on demand, so it updates each time you add the brand new knowledge.
Inform Claude what to incorporate
Immediate Claude to construct the visible dashboard related to your grasp sheet. Be sure to change [sheet name] together with your Sheet’s title and [domain] together with your area:
The freshness warnings cease you from presenting stale knowledge as present, and the “evaluate, by no means mix” instruction retains knowledge separate. The chance part in that immediate is how the dashboard makes suggestions based mostly on the information.
In your first construct, the prompt-tracking charts will present only some days of knowledge. That is anticipated.
On our check website, the primary refresh flagged three issues that went straight onto the content material plan. Step 8 walks by way of every one.
Step 7: Take a look at that the dashboard updates accurately
The entire level of the dashboard is to trace visibility over time, so that you’ll wish to be sure it updates correctly. Take a look at the connection by altering one quantity within the Sheet and confirming it strikes within the dashboard.
Edit a clicks worth within the “gsc_pages” tab, refresh the artifact, and verify that the matching KPI card adjustments. Test the data-freshness stamp exhibits the present learn, then undo the edit.
If a piece of the dashboard goes empty after a Sheet edit, evaluate your column headers in opposition to the unique construction. They should match precisely.
Step 8: Evaluate the dashboard and prioritize actions
Use the dashboard to resolve what to work on subsequent. This is how we learn the three findings from our check website’s first refresh:
- A web page AI cites that earns no Google clicks: Our executor guidelines was cited for 50 prompts and earned zero clicks. AI platforms already belief the web page, so the work is on the Google aspect. Discover a key phrase it may well realistically rank for, refresh the web page round it, and construct hyperlinks to it.
- Rivals cited the place you are not: Our competitor checklist was led by domains cited for seven prompts every. Open these prompts and skim what will get cited. That is your checklist of comparability and class content material to create, and of listicles to get talked about in.
- An entire immediate group with no presence: We had zero visibility throughout all six customer-problem prompts. One weak web page does not clarify that. It factors to lacking content material, so these six prompts turned the transient for brand new top-of-funnel items.
As your historical past builds, look ahead to these patterns too:
- A web page rating effectively in Google however hardly ever cited by AI often wants clearer construction: direct solutions, descriptive headings, and present dates that make passages simple to carry
- A web page dropping mentions after a content material change is your early warning to overview the edit
- Subjects bettering throughout each AI solutions and Google search are your proof of what works, so doc what these pages have in widespread
You’ll additionally see an “Alternatives” part, the place Claude evaluations the overlap between GSC and Semrush knowledge and provides you ideas on how one can enhance your AI visibility.
For instance, a web page can rank effectively in Google and nonetheless by no means seem in ChatGPT. ChatGPT pulls from its personal net index, so a powerful Google rating does not assure a quotation. Test whether or not the web page solutions the tracked immediate instantly, and whether or not the sources ChatGPT does cite for that immediate point out you. If Bing and Copilot matter to you, Semrush’s Place Monitoring device tracks Bing rankings, and you’ll add that knowledge to the dashboard with one additional export.
Step 9: Replace the dashboard every week
You’ll must replace the Sheet on a set weekly routine to maintain the information correct. Every week, ideally on the identical day:
- Export the newest immediate monitoring knowledge (or gather it from the scheduled export e-mail)
- Export the newest Cited Pages and Cited Sources tables from Visibility Overview
- Export the newest GSC knowledge, ending the window 2-3 days prior
- Append the brand new rows with the brand new snapshot date, then deliver the up to date file into the identical Google Sheet with File > Import > Exchange spreadsheet. In the event you import it as a brand new file, the dashboard can hold studying the previous one.
- Refresh the artifact and make sure the freshness stamps up to date
- Observe any main content material adjustments or campaigns within the Sheet, so subsequent month you’ll be able to join actions to causes
Whenever you replace the sheet, hold the following tips in thoughts:
- Hold the identical headers
- Stamp each Semrush export batch with its snapshot date
- Append new rows fairly than changing previous ones so that you get comparability knowledge
- Normalize URLs persistently throughout tabs
In the event you don’t normalize URLs, GSC can report a few of your pages twice; as soon as with a trailing slash and as soon as with out. In case your page-level numbers ever look absurdly low, URL variants are the very first thing to verify.
Construct your AI visibility dashboard
A model visibility dashboard for AI search exhibits you ways typically AI cites and mentions your model, subsequent to what those self same pages earn in Google.
Open the AI Visibility Toolkit, create your first monitoring marketing campaign, and write 30 prompts the best way your consumers really ask. Then join that knowledge together with your grasp sheet and construct the dashboard in Claude.
And if you want to present the dashboard to stakeholders, ask Claude to show the artifact right into a shareable HTML file with the identical knowledge.
