Buyers increasingly ask AI assistants for recommendations before they ever visit a website. A shortlist inside ChatGPT, a cited domain in Gemini, or a first mention in Grok can shape which brands get considered—and which never appear. That is the new discovery layer: AI visibility, not just crawl access or classic rankings.
Traditional SEO tools answer whether search engines can index your pages. Generative assistants answer a different question: when someone asks for the best option in your category, do major AI engines mention your company, mention you before competitors, cite your domain, and speak about you positively? AI visibility tracking is how you measure that—systematically, across engines, with repeatable runs.
The problem: recommendations happen before the click
Search still matters. So does technical SEO, content quality, and authority. But generative answers compress the journey. Users type a job-to-be-done, a category query, or a “best X for Y” prompt and receive a curated narrative: brands named, alternatives compared, sources cited, opinions framed.
If your business is absent from those answers—or appears after competitors, without a first-party citation, or with lukewarm sentiment—you lose mindshare before the website visit. That gap is hard to spot with conventional rank trackers. AI answers vary by engine, prompt, and time. Measuring AI brand mentions, first mentions, citations, and sentiment over repeated runs is a different discipline than tracking a blue-link position.
Teams already experiment with screenshots and one-off prompts. That does not scale. You need structured ChatGPT brand monitoring, Gemini citation tracking, and Grok visibility on the same prompt packs—plus history so changes are visible after enough runs. The goal is operational: collect, score, compare, export, improve content, and collect again—treating AI answers as a measurable discovery surface rather than a black box.
What AI Visibility is (and is not)
AI Visibility from PositionMySite (also known as SEO Site Signals / tools.positionmysite.com) shows whether major AI assistants mention, recommend, and cite a business. It turns live AI answers into measurable visibility, competitor intel, and actionable content recommendations.
It is not a conventional rank tracker. It does not claim identical or forever-reproducible answers. It does not guarantee rankings. And it is not merely a check that AI crawlers can access your site—that is a related but separate concern (e.g. llms.txt and crawl access). The primary angle here is measuring live answer mentions, citations, and sentiment with bring-your-own-key (BYOK) multi-engine collection.
Access is limited to paid subscribers and complimentary accounts. You must sign in. The product is not free or anonymous. Every user brings their own API key from at least one supported provider. PositionMySite does not pay provider API charges and does not silently substitute its own key.
How the workflow works
The collection model is straightforward:
- Select a website and a prompt pack.
- Run ten live, web-enabled prompts against the engines you have keyed.
- Review whether engines mention your company, mention you before competitors, cite your domain, and express positive, neutral, or negative sentiment.
- Compare engines on the same prompts, flag model disagreements, and inspect competing brands and share of answers.
- Use content-gap findings and recommended pages to create or strengthen, then export CSV or a structured JSON brief for further planning in ChatGPT, Grok, Claude, Gemini, or Perplexity.
Two prompt packs cover complementary angles:
- Market pack — category, competitor, geo, pricing, and jobs-to-be-done (JTBD) prompts.
- Brand pack — reviews, pricing, alternatives, opinions, and direct brand questions.
Together they support AI search optimization and generative engine optimization (GEO) workflows: you see what assistants say today, where you are missing, and what to publish or reinforce next.
Live collection currently covers Grok (xAI), ChatGPT (OpenAI), and Gemini (Google). One key is enough to start; more engines mean broader comparisons. Claude and Perplexity are not live-collected—they can consume the exported JSON brief for content planning only.
Each collection run applies to the currently selected website only. Weekly collection is typically enough to see trends. Charts become more meaningful after three runs with the same pack.
BYOK, keys, and security
AI Visibility is built around your keys, your quota, and clear scoping:
- Keys are validated before save.
- Keys are encrypted after submit.
- The UI shows only the last four characters.
- You can replace or remove keys at any time.
- Keys can scope to one site, selected sites, or all sites.
- Collections use your provider quota—not an unlimited PositionMySite-funded pool.
The dashboard estimates API cost before collection. Estimates update based on selected engines and sites. Because provider prices change, treat estimates as guidance—not permanent published dollar claims.
This BYOK model keeps cost transparent and under your control while enabling real multi-engine AI citation monitoring and AI search competitor analysis.
How the AI visibility score works
Each run produces a score weighted across four signals:
- Mention — 40% — Does the engine mention your company?
- First-mention — 25% — Are you mentioned before competitors?
- First-party domain citations — 20% — Does the answer cite your domain?
- Sentiment — 15% — Is tone positive, neutral, or negative?
Beyond the score, the product surfaces competing brands and share of answers, flags model disagreements across engines, and helps separate your real domain from directories, name-twins, and unrelated sources. That last piece matters for accurate GEO reporting: not every citation that looks similar is yours.
Dashboard: Overview, Prompts, Citations, Sentiment, History
The AI Visibility dashboard is organized for diagnosis and iteration:
- Overview — high-level visibility across engines and packs.
- Prompts — the ten live prompts and how answers behave.
- Citations — where engines point, including first-party vs. other sources.
- Sentiment — positive / neutral / negative framing over answers.
- History — change across runs so you can see movement after repeated collection.
Exports include CSV plus a structured JSON brief designed for content planning in ChatGPT, Grok, Claude, Gemini, and Perplexity—even where Claude and Perplexity are not part of live collection. That split matters for honest AI search optimization workflows: live measurement on the engines you have keyed, then planning assistance wherever your team already writes.
Because each collection covers only the currently selected website, multi-site teams should expect to run per property. That keeps cost estimates and quota usage tied to intentional scope rather than a silent bulk pull across every property at once.
Who it is for — use cases
Business owners want a clear answer: do assistants mention us when someone asks for options in our category? AI Visibility turns that into a measured baseline, not a guessing game. Instead of relying on a handful of manual prompts, you get a consistent pack, a score you can explain, and history that shows whether visibility is improving after content work.
SEOs and marketing teams need AI search competitor analysis alongside classic SERP work. Mentions, first mentions, citations, and sentiment give a parallel funnel of visibility—useful for content roadmaps and executive reporting. Pair Market-pack category and JTBD prompts with Brand-pack review and alternatives prompts to see both “who wins the category narrative” and “what the engines say when asked about you directly.”
Agencies can run packs for client sites, compare engines, export briefs, and recommend pages to create or strengthen—without pretending AI answers are static rankings. Client reporting can show engine-level differences, flagged disagreements, and citation quality (first-party domain vs. directories or name-twins), which keeps the conversation grounded in what was actually observed in live answers.
SaaS and category-focused brands often care most about share of answers vs. named competitors. Tracking competing brands and model disagreements helps prioritize messaging and proof pages that assistants (and humans) can cite. When one engine cites you and another does not, that disagreement is a signal to investigate content coverage, entity clarity, and third-party sources—not a reason to assume a fixed “rank.”
Common workflows include:
- Baseline Market + Brand packs, then weekly re-runs on the same pack.
- Engine comparison when answers diverge (e.g. ChatGPT vs. Gemini citations).
- Content-gap follow-through: create or strengthen recommended pages, then re-measure.
- Export JSON briefs into your preferred writing stack for planning—not for claiming live Claude/Perplexity collection.
- Scope API keys carefully (one site, selected sites, or all sites) so collections stay aligned with how teams manage provider quota.
Positioning for stakeholders is straightforward: this is generative engine optimization measurement and GEO reporting support—visibility inside answers—not a promise that every assistant will recommend you tomorrow.
FAQ
Is AI Visibility free?
No. Access is for paid subscribers and complimentary accounts only. You must sign in. Every user brings their own API key from at least one supported provider. PositionMySite does not pay provider API charges.
Which engines are live-collected?
Grok (xAI), ChatGPT (OpenAI), and Gemini (Google). Claude and Perplexity are not live-collected; they can use the exported JSON brief for content planning.
Do I need keys for every engine?
One key is enough to start. More engines enable broader comparisons on the same prompts.
How are my API keys handled?
Validated before save, encrypted after submit, UI shows last four characters only. You can replace or remove keys. Scope keys to one site, selected sites, or all sites. Collections use your provider quota.
How often should I collect?
Weekly collection is typically enough for trends. Charts become more meaningful after three runs with the same pack. Each collection covers the currently selected website only.
Will this guarantee rankings or identical AI answers?
No. AI answers vary. This product measures mentions, first mentions, citations, sentiment, and change over repeats—not conventional rank positions or forever-reproducible outputs.
Is this the same as checking if AI crawlers can access my site?
No. Crawl access and files like llms.txt matter, but AI Visibility focuses on whether live assistant answers mention, cite, and recommend your brand—with multi-engine BYOK collection.
What can I export?
CSV and a structured JSON brief for content planning across ChatGPT, Grok, Claude, Gemini, and Perplexity.
Start measuring AI visibility
If recommendations inside AI assistants are becoming part of how buyers shortlist vendors, you need more than anecdotal prompts. You need structured AI visibility tracking, AI brand mentions analysis, and AI citation monitoring across the engines that matter—with transparent BYOK costs and history you can act on.
Canonical CTA: open the AI Visibility dashboard at https://tools.positionmysite.com/dashboard/ai-visibility (sign-in required; paid subscribers and complimentary accounts; bring your own API key).
PositionMySite / SEO Site Signals — tools.positionmysite.com