AI Visibility Documentation

AI Visibility tracks how often — and how well — your brand is mentioned when people ask ChatGPT, Gemini, Perplexity, and Google AI Mode for recommendations in your category. Traditional SEO tells you how you rank on Google. AI Visibility tells you whether you exist at all inside an AI-generated answer.

A concrete example

Say you sell project-management software. Someone opens ChatGPT and types “what's the best project management tool for a 10-person startup?”. The AI doesn't run a Google search behind the scenes and show ten blue links — it composes one answer, usually naming two or three tools. If your brand is one of them, you just won a customer without them ever visiting your website first. If it isn't, you were never in the running, and you'd have no way of knowing that happened at all — unless you were tracking it. That's exactly the gap AI Visibility fills.

Why it matters

A growing share of buyers now ask an AI assistant instead of typing a search query, especially for comparison and “best of” style questions where they want a synthesized recommendation rather than a list of pages to research themselves. If your brand isn't surfacing in those answers, you're invisible to that audience — no matter how well you rank on Google, and no matter how good your product actually is. Worse, you can't fix a problem you can't see: without a tool like this, most teams find out they're missing from AI answers only when a customer mentions it in passing, if ever. AI Visibility turns that blind spot into a measurable, actionable dashboard you can check the same way you'd check Google Analytics.

How this is different from traditional SEO

• SEO measures rank — your position in a list of links. AI Visibility measures presence — whether you're mentioned inside a generated answer at all, and where in that answer.
• SEO tools crawl and score your own pages. AI Visibility actually asks the AI engines your buyers' real questions and reads what they say back — it's observing the AI's behavior directly, not inferring it from page signals.
• A page can rank #1 on Google and still never get mentioned by an AI engine, because AI answers draw on a different mix of signals (training data, citations, brand authority across the web) than classic ranking factors.
• SEO is a single-channel game (Google). AI Visibility is inherently multi-engine, because your buyers don't all use the same assistant.

What you get

Once a project is set up and scanning, your dashboard turns raw AI answers into five kinds of insight:

• A 0–100 AI Visibility Score, tracked day over day, so you can see whether you're trending up or down at a glance.
• Mention, sentiment, and citation tracking across every engine you scan — not just whether you're named, but how positively, and what sources the AI is trusting.
• Side-by-side comparisons against every competitor you track, prompt by prompt and engine by engine.
• Opportunity gaps — the exact list of questions where a competitor is winning and you're not, which is usually the fastest place to start improving.
• A prioritized AEO/GEO action plan that turns all of the above into a concrete list of content and outreach ideas, so you're never just staring at a score with no idea what to do about it.

A project is one brand/domain you want to track — most accounts only ever need one, but agencies or multi-brand companies can run a separate project per brand. Setup is a guided 4-step wizard at /dashboard/ai-visibility/create, and takes a few minutes end to end.

1. 1Project details — plan, brand, region.
2. 2Review the prompts AI Visibility generated for you.
3. 3Add competitors to benchmark against.
4. 4Preview your dashboard and subscribe.

Step 1 — Project details

This step tells AI Visibility who you are and where to look for you.

• Pick a plan (Starter / Pro / Growth) using the segmented control at the top — this decides how many prompts you can track per day and how many competitors you can add. See Plans & Billing for the full breakdown; you can always contact support later to move to a different tier.
• Enter your brand's website (e.g. yourcompany.com). AI Visibility auto-detects your brand name from the site so you don't have to type it — you can edit it, and add extra alias names (e.g. a short form, an old name, or how people casually refer to you). The first name in the list is treated as the primary name for matching mentions.
• Choose your target country and language. AI engines tailor their answers to the asker's region and language, so a scan run “as a US English speaker” can look completely different from one run “as a German speaker in Germany” — pick whichever matches where your actual buyers are.
• Optionally connect Google Analytics & Search Console in one click. This isn't required to run AI Visibility, but it gives you richer context to cross-reference against (e.g. is a jump in AI mentions showing up as a jump in real traffic too?).

Step 2 — Review prompts

This is the heart of setup: AI Visibility crawls your site and generates a realistic set of prompts — the actual questions real buyers would type into ChatGPT or Perplexity — grouped into categories such as discovery (“best tools for X”), comparison (“X vs Y”), use-case (“tool for [specific job]”), problem-solving, and best-of lists, among others.

• Review the generated list and remove anything irrelevant to your business — the goal is a set of prompts your actual buyers would plausibly type, not every prompt the generator could think of.
• Add your own custom prompts for anything specific you want tracked that wasn't auto-generated — a niche use case, a seasonal question, or a phrase your sales team hears often.
• The prompts you keep here become the pool that later gets split into Core, Control, and Rotating buckets once scanning starts — see Running & Scheduling Scans for exactly how that works.

Step 3 — Competitors

AI Visibility suggests competitor domains based on what it found while crawling your site and market. Add or remove competitors as needed — you don't have to use the suggestions, and you can add competitors you know about that weren't suggested.

• Give a competitor alias names if they're commonly referred to by more than one name (a shortened brand name, an old name post-rebrand, etc.) — this makes mention-matching more accurate.
• The number of competitors you can track is capped by your plan tier (5 on Starter, 10 on Pro, 15 on Growth) — pick the ones that matter most if you're near the limit; you can always swap competitors later from the project's Edit page.

Step 4 — Preview & subscribe

You'll see a preview of what your dashboard will look like once real data comes in, along with your monthly price based on the plan you picked. Clicking Subscribe takes you to a secure hosted checkout page; once payment completes, you're redirected back, your project is actually created in the system, and you're ready to run your first scan.

A scan is the actual work of AI Visibility: it fires every one of that day's tracked prompts at every tracked engine, reads the response, and scores whether your brand was mentioned, how it was framed, and what sources the AI cited — then stores all of it so every dashboard tab has fresh data to show. A full scan typically takes 3–5 minutes; you'll get an email the moment your report is ready, so there's no need to sit and watch it run.

Automatic daily scans

Every active project scans automatically once a day at 02:30 UTC — this is a fixed schedule you don't need to configure. Both the project list page and the project dashboard show a live countdown to the next scheduled scan, so you always know when your data will next refresh.

Your first scan is manual

Because there's no historical data to schedule around yet, your very first scan starts with a “Start First Scan” button on your project card. After that one kicks off, every following scan runs on the daily schedule automatically — no action needed from you unless you want to jump in and check on something sooner.

Prompt buckets — how coverage works day to day

With a large prompt list and a plan tier that only allows a limited number of scans per day, AI Visibility can't realistically run every single prompt, every single day forever — so prompts are split into three buckets that balance stable trend-tracking against broad topic coverage:

• Core prompts — a fixed set that runs every single day without exception. These are your most important, highest-intent questions — the ones you'd want a clean, uninterrupted trend line for.
• Control prompts — another fixed daily set, run purely as a stability baseline. Because these never change, any movement here tells you whether the AI engines themselves are behaving differently (model updates, general drift) rather than something specific to your brand.
• Rotating pool — everything else, cycled through a schedule of up to 30 days. Instead of scanning all of it daily, AI Visibility scans a slice each day and rotates through the full pool over the cycle — so over a month you still get complete coverage of every prompt you added, just spread out sensibly.

In practice this means: your Overview score and trend charts update daily from Core + Control, while your Results and Prompt Coverage numbers fill in more completely as the Rotating pool cycles through — both views stay meaningful, just on slightly different rhythms.

The first thing you see when you open a project — the single screen that tells you where you stand right now, without digging into any other tab. When the most recently scanned day is selected, this tab relabels itself “Today” so it always reads naturally.

What's on this screen

• A circular AI Visibility Score gauge (0–100, rated Strong / Growing / Low) plus a mentions-by-LLM breakdown showing which engines are talking about you most.
• A Prompt Coverage donut — how many of your tracked prompts you're visible in versus gaps, at a glance.
• A 6-metric strip: Mentions, Share of Voice, Queries Scanned, Prompt Gaps, Sentiment, and the selected Scan Date/Range.
• Distribution by LLM, a competitor Share of Voice comparison, and two trend charts — overall visibility and per-engine visibility across your recent scan days.

How to read it

Treat the Score gauge as your headline number and the trend chart as the more important signal — a score of 60 that's been climbing for two weeks is a much better story than a score of 75 that's flat or sliding. If the Score drops sharply on one day, check whether Queries Scanned also dropped (a scan issue) before assuming your actual visibility changed. Use the date-range picker in the sidebar to compare “today” against last week or last month rather than judging a single day in isolation.

Your prioritized growth roadmap. “AEO/GEO” stands for Answer Engine Optimization / Generative Engine Optimization — the practice of optimizing content so AI engines cite and recommend you. This tab is disabled until your first scan completes.

Instead of asking a single AI model to grade its own homework, AI Visibility sends your full result set to four different LLMs, has each one independently analyze it, and then has a judge model reconcile their opinions into one coherent report — this cross-checking is what keeps the recommendations grounded in what your data actually shows, rather than one model's guess. The report reads top to bottom in this order:

• Executive Summary — the headline read on where you stand and what to do first, written so you could hand it to a manager with no other context.
• Marketing Calendar — concrete content ideas by channel: blog posts, comparison articles, YouTube/TikTok/Reels, LinkedIn, Instagram/Facebook, Reddit/Quora, Wikipedia edits, keyword targets, influencer & partner ideas, PR/backlink outreach, review-site listings, and more. Each idea exists because it targets a specific gap or opportunity found in your actual scan data — nothing here is generic advice.
• Roadmap — the same ideas laid out on a phased timeline, so you know what to tackle in weeks 1–2 versus what can wait until month two.
• Take Action — a projected-impact snapshot plus a prioritized (P1/P2/P3) action table covering page-level fixes, AI-specific optimizations, and content/keyword focus areas. Mark each item To Do, In Progress, or Resolved as you work through it — your progress is saved per project.
• More Insights — supporting analysis on your market positioning, brand monitoring, and how your GEO performance is trending over time.

How to use it

Don't try to do everything at once. Start with the P1 items in Take Action — those are flagged because the judge models found them to be the highest-impact, lowest-effort fixes given your current gaps. As you publish content or make changes from the plan, mark items In Progress so the next scan's Opportunities tab becomes the real test of whether it worked.

Overview tells you whether you're mentioned. Brand tells you how well — how each AI engine actually perceives your brand relative to competitors, not just whether you show up, but how prominently and how often within a single answer.

What's on this screen

• Three headline stats: AI Visibility Score, Share of Voice, and Brand Sentiment.
• Share of Voice by LLM — your mentions vs. every tracked competitor, broken down per engine, so you can spot if you do great on ChatGPT but poorly on Perplexity, for example.
• Who Gets Named First? — how often you're the 1st, 2nd, 3rd+, or not-mentioned brand in an answer. Being named first matters a lot in practice: many users only read the first recommendation an AI gives them.
• How Much Does AI Talk About You? — mention-frequency distribution within a single answer (3+ times, twice, once, or not at all) — a proxy for how central you are to the AI's framing of the topic, versus a passing mention.
• Where You Win — First Mention — the exact prompts where you're named ahead of every competitor, useful as proof points for sales or marketing copy.
• Competitor Dominance Matrix — a grid of prompts × brands, color-coded win / shared / loss / quiet, so you can scan an entire competitive landscape in one glance.

How to read it

Share of Voice is the single best number for “are we winning or losing” because it's relative — a rising visibility score means nothing if your competitors are rising faster. Use Who Gets Named First alongside it: a brand can have decent Share of Voice but rarely be the first name mentioned, which is a different (and often more fixable) problem than not being mentioned at all.

If Brand tells you how you compare, Opportunities tells you exactly what to do next: every prompt where a competitor showed up and you didn't — each row here is a specific, actionable content gap you can close to grow your visibility.

What's on this screen

• A summary strip: Gap Prompts, Prompts Visible, and Coverage Rate — your at-a-glance “how much work is left” number.
• Opportunity Gaps table — which engines a prompt is missing from, and which competitors were mentioned instead; expand any row to read the full AI response and see exactly how the AI framed the answer without you in it.
• Brand Visible Prompts table — prompts where you ARE mentioned, with sentiment, your best position, and which competitors also showed up alongside you (useful to confirm you're not just visible, but visible favorably).

How to use it

Work top-down and treat each row as a mini brief: read what the AI actually said for that prompt, note which competitor(s) it named instead, and ask what content or evidence would make a reasonable AI cite you for that exact question. This is also the fastest way to sanity-check the AEO/GEO Action Plan — a lot of its recommendations are literally generated by working through this exact table for you.

Where Brand and Opportunities focus on you, Competitors flips the lens — a full, side-by-side visibility comparison for every competitor you track, engine by engine.

What's on this screen

• A ranked Competitor Visibility Comparison leaderboard — every tracked brand ordered by visibility score, with the option to show or hide any competitor from the view to declutter the comparison.
• Click any competitor for a head-to-head deep-dive: visibility comparison, prompts where only they appear, prompts where only you appear, per-engine mention counts, and a full prompt-level table of every place they show up.

How to use it

The “prompts where only they appear” list in the deep-dive is the single most useful view here — it's a distilled version of Opportunities filtered to exactly one rival, which is often exactly what you need when you're specifically trying to catch up to (or stay ahead of) one named competitor rather than the whole field at once.

Every other tab analyzes prompts you've already chosen to track. Prompt Research is where you find the next ones — real questions people are actually asking AI in your category, discovered before you commit to tracking them.

How it works

• Search by topic, brand, or domain, and choose a platform lens (Google AI or ChatGPT) to see the kind of questions asked on that surface.
• Results include estimated monthly AI search volume, a category tag (discovery, comparison, use-case, problem-solving, best-of), and a mini trend line showing whether interest in that question is rising or falling.
• Add any suggestion straight into your tracked prompts with one click — it joins your Rotating pool immediately (see Running & Scheduling Scans).

How to use it

Use this before you assume you know every question your buyers ask — it's common to discover high-volume questions here that never would have made it onto a prompt list written from memory. Prioritize adding suggestions with both meaningful volume and a rising trend line; a niche, flat-volume question is lower priority than one that's clearly becoming more common.

Every other tab is a curated view built on top of one thing: the raw, prompt-by-prompt record of every scan. Results is that ground truth, unfiltered — the table every score and chart elsewhere on the dashboard is ultimately computed from.

What's on this screen

• Search, filter by mention status (All / Mentioned / Not Mentioned), and filter by one or more engines — useful when you only care about a specific slice, like “everything Perplexity said this week.”
• Each row shows a 0–100 visibility score, a VISIBLE / NOT MENTIONED badge, per-engine indicator dots, and the top competing brands also mentioned in that same response.
• Click any row to open the full response text from each engine — exactly what the AI said, word for word, with no summarizing or editing in between.

When to use this instead of another tab

Reach for Results whenever a summary number surprises you and you want to verify it against the actual source. If your Sentiment score drops unexpectedly, or a Brand metric looks off, this is where you go read the exact response text that produced it — nothing here is inferred or aggregated away.

Appearing in an answer isn't enough on its own — an AI can mention your brand while comparing you unfavorably to a rival, or flagging a downside. Sentiment tells you how you're framed, not just whether you're present.

What's on this screen

• Positive / Neutral / Negative mention percentages at a glance — your overall tone breakdown across every tracked prompt.
• Sentiment by LLM — a stacked bar per engine, since different engines can frame the same brand differently based on what they've read about you.
• Competitor Sentiment Comparison — a −1 to +1 sentiment score, overall and per engine, for you and every competitor, so you know if a rival isn't just more visible but more favorably discussed.
• Prompts by Sentiment — an expandable list grouped by tone, linking straight to the full response text so you can read exactly what triggered a negative or neutral read.

How to read it

A high visibility score paired with weak or negative sentiment is a real warning sign — it usually means you're being mentioned mainly in a comparison context where you come out worse, which is worth investigating in the underlying response text before you invest in more content aimed at pure visibility.

Some AI engines don't just compose an answer from memory — they actively search the web and cite sources, the same way a person would open a few tabs to research an answer. Citations shows you every domain those engines actually trusted enough to reference in your niche, ranked by how much influence that source carries.

What's on this screen

• Filter by engine — citations mainly come from web-search-enabled engines like Perplexity and Google AI Mode, since not every engine cites sources the same way.
• Three headline stats: your own Domain Influence Score, the share of your brand-mentioning responses that include a citation, and the same figure for non-brand responses.
• A Top Domains treemap sized by citation frequency, plus a full table of every cited domain with its Influence Score, source category, prompt count, and a list of the exact links cited.

Why this matters

This tab effectively reveals which third-party sites AI engines treat as authoritative in your category — review sites, forums, comparison articles, and news outlets that keep showing up as citations. Getting your brand listed, reviewed, or covered on those exact domains is one of the most direct ways to influence what an AI engine says about you, because you're showing up in the sources it already trusts rather than trying to get the AI to trust a brand-new source.

Every other tab shows you your data. This one shows you the machinery behind it — your full scan history, period reports, and project housekeeping. It's labeled “Runs” internally, but renamed “Settings & Reports” in the sidebar since that better describes what you'll actually use it for.

Scan history (Daily)

A complete log of every scan ever run for this project: date/time, whether it was triggered manually or by the daily schedule, its status, how many prompt × engine queries completed, and quick actions to view that day's report, download it as a PDF, or share it. This is your first stop if a day's data looks off — check whether that day's scan actually completed successfully before assuming a real change happened.

Weekly & Monthly period reports

Switch to Weekly or Monthly, pick a start date, and generate a consolidated report over that window — trend stats, engine and competitor trends, and a rolled-up action plan spanning the whole period instead of just one day. This is useful for anything you'd otherwise have to piece together by comparing several Daily Reports by hand — a weekly summary for a team standup, or a monthly report for a client or stakeholder. Once generated, a period is locked in the calendar so you won't accidentally regenerate (and pay the generation cost for) the same window twice.

Billing

A card here links straight to your subscription and invoices under Settings → AI Visibility — see Plans & Billing for the full picture of how that works.

Danger zone

Delete a project here when you're done with it. This permanently removes all of its run history and results, so make sure you've downloaded or shared anything you want to keep first. Deleting a project with an active subscription automatically cancels that subscription too — you won't need to cancel it separately beforehand.

Your setup isn't locked in forever. Open “Manage Project” from your project's sidebar at any time to adjust who you track and how — most changes take effect from the next scan onward.

What you can change

• Brand name(s) — add or remove alias names, for example if you rebrand, shorten your name in casual use, or acquire a product under a different name. The first name in the list is treated as primary for mention-matching, so put your main brand name first.
• Competitors — add or remove domains, each with optional alias names, up to your plan's competitor limit (a live counter shows how many you've used). Good times to revisit this: a new competitor enters your market, or one you're tracking is no longer relevant.
• Country & language — change the region your prompts are evaluated for, useful if your target market shifts (e.g. you expand from a US-only audience into UK or EU buyers).
• Prompt buckets — Core, Control, and Rotating prompts are shown here for reference. The one direct action available is promoting a Rotating prompt into permanent Core coverage — do this for any prompt that turns out to matter more than expected once you start seeing real data.

What you can't change here

Your brand's domain and plan tier aren't editable from this page by design — the domain is your project's identity, and plan tier changes affect billing, so both are handled through support instead of a self-service form. If you need to change either, reach out and the team will help directly.

The interactive dashboard (Overview, Brand, Opportunities, and the rest) is built for exploring and comparing data across days. But sometimes you just want one clean, fixed document to read top-to-bottom, print, or hand to someone else — that's what the Daily Report is for: a standalone, linear snapshot of one specific scan.

Verdict tags — the headline at a glance

Every report opens with a single-word verdict summarizing that run compared to the one before it, so you don't have to read the whole document just to know if things are going well:

• Rise — your visibility improved since the last scan.
• Drop — your visibility declined since the last scan.
• Lead — you're ahead of your tracked competitors this run.
• Flat — little to no meaningful change.
• First Scan — this is the project's very first run, so there's no prior data to compare against yet.

What's inside

Sections run top to bottom in a deliberate order: executive summary, key stats, visibility score & rank, share-of-voice, sentiment (overall and by engine), visibility by engine, a competitor matrix, prompt coverage, a trend chart, opportunity gaps, top mentions, cited domains, and finally the full, untrimmed results for every single prompt — so a reader who only has five minutes gets the summary up top, and anyone who wants to verify a claim can keep scrolling to the raw evidence.

Downloading and sharing

• Download PDF — export the report exactly as it appears, to share internally, attach to an email, or archive outside the dashboard.
• Share — generate a public, read-only link that anyone can open without logging in. This is perfect for sending a snapshot to a client, an executive, or a stakeholder who doesn't have (and doesn't need) a dashboard account — they see exactly the report as it was at the time you shared it.

Weekly & Monthly reports

A Daily Report always covers one scan. If you want the same kind of document but rolled up over a longer window, generate a Weekly or Monthly period report instead from the Settings & Reports tab — see that section for details.

Every plan tracks the same four base engines — ChatGPT, Gemini, Perplexity, and Google AI Mode. What changes between tiers is scale: how many prompts you can scan per day, and how many competitors you can track alongside your own brand.

Choosing a tier

As a rough guide: Starter suits a single brand tracking a focused set of core questions and a handful of direct competitors. Growth (the default) fits most established brands that want broader prompt coverage and a fuller competitive set. Pro is built for brands in a crowded category who need a large competitor field alongside solid prompt coverage. If you're unsure, Growth is a safe middle ground — you can always move up later.

Claude add-on

Claude tracking is available as a separate, per-engine add-on on top of any plan — billed and managed independently of your base subscription. It's worth adding if a meaningful share of your buyers skew technical or enterprise, since Claude tends to be the assistant of choice in those circles even when it isn't the most-used engine overall.

How billing works

• Each project has one base subscription and, optionally, its own Claude add-on subscription — each with independent status, price, and renewal date, so you can add or drop the Claude add-on without touching your base plan.
• Cancelling your base subscription stops auto-renew: you keep full access until the current billing period ends, and only then are the subscription and all project data removed. No refunds are issued for the remainder of the current period, so it's worth timing a cancellation close to your renewal date if you're trying to make the most of what you've paid for.
• Cancelling the Claude add-on works the same way — access continues until the period ends, no refund, and your base subscription (and the four core engines) is unaffected.
• Deleting a project with an active subscription cancels that subscription automatically — you don't need to cancel it separately before deleting.

Invoices

Find every subscription and downloadable payment invoice under Settings → AI Visibility (also linked from each project's Settings & Reports tab). This view covers the base subscription and any add-ons for every project on your account, so it's the one place to go for a full billing history rather than checking project by project.

Every AI Visibility–specific term used across this documentation and the dashboard itself, defined plainly — with a quick example wherever a plain definition alone might still leave you guessing.