Built in your accounts · included in the program
AI Search Reporting: Rankings, AI Answers and Signups in One Place
We build a reporting system in your own accounts that pulls Search Console, Bing, Google Analytics, your rank tracker and your AI answer tracker every morning, so you can see every page and every buyer prompt in one place. It tracks whether ChatGPT, Claude, Perplexity and Gemini name you, pick you and cite you, and ties it all to signups. Numbers come from scripts, never from AI. Priced after a call, and included in our AI Search Program.
The short version
What is AI search reporting?
AI search reporting tracks how a company shows up in Google and in AI assistants, side by side, and ties both to signups. It combines classic SEO data (clicks, impressions and positions per URL) with AI visibility data (whether ChatGPT, Claude, Perplexity and Gemini name the brand, recommend it and cite its pages for each buyer question), refreshed daily so you can see which changes moved which numbers.
SuperDupr builds it in your accounts and runs it as part of the AI Search Program, or on its own, priced after a call.
Where does the data come from?
Five sources and your own site, pulled by scripts every morning. Each number keeps a note of where it came from.
| Source | What it gives us | What to know |
|---|---|---|
| Google Search Console | Clicks, impressions and average position per URL and query | Google counts AI Overviews and AI Mode inside the regular Web numbers, so they aren't split out |
| Bing Webmaster Tools | Clicks and positions in Bing, plus AI citation data | Bing's AI Performance report (public preview since February 2026) shows citations in Copilot and Bing AI summaries by page |
| Google Analytics | Sessions and signups by landing page and source | Signups need a conversion event set up; we check it first |
| Rank tracker (e.g. Ahrefs) | Daily positions for the keywords you care about, and competitors | Fills the gaps Search Console leaves for low-volume terms |
| AI answer tracker (e.g. Scrunch or similar) | The answer each assistant gives to each tracked prompt | Answers vary run to run, so we read trends, not single days |
| Your site | Page score, last real update, word count, schema present | Pulled from the pages themselves, not from a CMS field |
Choosing an AI answer tracker? Justin compared the options in the best AI visibility tools.
Why two tables, and why no AI in the numbers?
Everything lands in two tables. The first has one row per URL per day: clicks, position, AI citations, page score, last update and signups. The second has one row per prompt per AI assistant per day: were you named, were you picked, and which sources were cited. Two simple tables mean any question about your search performance is one filter away, and any dashboard can be rebuilt from them.
The numbers come from scripts, never from AI. A model asked to summarize a spreadsheet can round, guess or invent a figure, and you'd have no way to tell. So scripts pull and store every number, and Claude only writes the short notes on what moved and why. A person reads those notes before they go out.
Why does "mentioned" differ from "picked"?
When a buyer asks an assistant which tool to use, the answer often lists several options and then recommends one. A mention rate counts you whenever you appear in the list. It can climb while a competitor gets recommended every time. We track three separate things, because each one points to a different fix.
| Signal | What it means | What usually fixes it |
|---|---|---|
| Named | You appear somewhere in the answer | Being on the pages assistants read: your own, review sites and lists |
| Picked | The assistant recommends you for the buyer's situation | A clear page for that audience or comparison, with facts that match everywhere |
| Cited | The assistant links a source to back its answer | Answer-first pages with real facts, so your page is the one worth quoting |
What does the SEO reporting dashboard show?
Six views, all built from the same two tables.
Every URL
Clicks, position, 7- and 28-day change, AI citations, page score, last update and signups, for every page on the site.
AI answers
Every tracked prompt by assistant: were you named, were you picked, and which sources were cited.
Movers
Pages and prompts that went up or down, each with a one-line reason.
Competitors
Who gets picked instead of you, new and changed competitor pages, and searches where a rival moved ahead.
Content pipeline
Pages being drafted, refreshed, merged or redirected, and how each one did after it shipped.
Fact mismatches
Places where a price, plan name or claim differs between your pages, third-party profiles and what assistants say.
How often will we hear from you?
Only when there's something worth reading. Nobody needs a daily email that says nothing changed.
| When | What you get | Where |
|---|---|---|
| Daily | An alert only when something moves: a page drops, a prompt flips to a competitor, a fact goes out of sync | Slack or email |
| Every Monday | A note of five lines or fewer: what moved, why, and what we're doing about it | Slack or email |
| Monthly | A report tied to signups: which pages and prompts brought them, and what's next | A shared doc and a short call |
Who builds it, and what can't it tell you?
Justin McKelvey builds and runs it. He's spent 15 years building products and shipped more than 50 of them. Before software, he ran snowboard and mountain-bike shops and taught himself to code. More about Justin.
We use it on our own sites. On our November 2, 2026 scan of each site's articles against our page standard, justinmckelvey.com scored 93 out of 100, superdupr.com 89 and getlocalcall.com 87. That page score is one of the columns in the URL table.
The honest limit. AI assistants don't give the same answer twice, and AI answer trackers sample prompts rather than see every real conversation, so we read trends over weeks, not single days. Search Console's latest days are still provisional; Google's API docs say recent values may still change, so the scripts re-pull them. Reporting shows what moved. It doesn't move anything by itself, and we don't promise rankings or citations. AI answers often shift within weeks of a page going live, Google takes months, and we measure from a week-2 baseline.
Where does reporting fit with the rest of the work?
Reporting is how the AI Search Program decides what to work on next, so it's included there. If you want a one-time diagnosis first, start with an SEO and AI search audit, or the $1,500 AI Search Audit for smaller sites. After a fixed AI Search Optimization build, $500 a month monitoring re-runs your prompts every month.
Planning a domain move? Daily reporting is how you catch a drop in the first week. See SEO migration services. On WordPress? See WordPress SEO services. Software company? See B2B SaaS SEO or our AEO agency page. Want a quick read on where you stand? Try the free AI visibility check.
Questions
- What is AI visibility tracking?
- AI visibility tracking records how AI assistants like ChatGPT, Claude, Perplexity and Gemini answer the questions your buyers ask, on a schedule. For each prompt and each assistant it notes whether your brand was named, whether it was the one recommended, and which sources were cited. Tracked daily, it shows which pages and fixes change the answers over time.
- Can you track AI Overviews in Google Search Console?
- Partly. Google counts AI Overview and AI Mode appearances in the regular Search Console Performance report under the Web search type, so the clicks and impressions are there but mixed in with everything else. To see which prompts show an AI answer and whether you're cited in it, you need a rank tracker or AI answer tracker on top, which is what this system joins together.
- What's the difference between being mentioned and being picked?
- Mentioned means the assistant named you somewhere in the answer, often in a list of five or six options. Picked means it recommended you as the answer to the buyer's question. A mention rate on its own can look healthy while a competitor wins every recommendation. We track both, plus the sources cited, because the pick is what drives signups.
- Which tools does the reporting use?
- Google Search Console, Bing Webmaster Tools, Google Analytics, a rank tracker such as Ahrefs, and an AI answer tracker such as Scrunch or a similar tool. If you already pay for these, we use your accounts. If you don't, we help you pick, and the scripts can be pointed at a different tool later.
- Do you use AI to write the reports?
- Only the notes. Every number comes from a script that pulls it from the source and stores it, so any figure can be traced back. Claude reads those tables and writes the short what-moved-and-why notes, and a person checks them before they go out. AI never generates or estimates a number in the report.
- How much does AI search reporting cost?
- It's priced after a 30-minute call, because it depends on how many pages and prompts you track and which tools you already have. It's included in our AI Search Program. You pay the tool vendors directly for any subscriptions, and everything is built in your accounts.
- Who owns the data and the dashboard?
- You do. The scripts, tables and dashboards are set up in your own accounts, so if we stop working together, everything keeps running where it is.
See what your reporting would show
On a 30-minute call we'll look at your current reports and a few AI answers together. You'll leave knowing which numbers you're missing and how we'd get them.