FAQ

Frequently asked questions

What does the free audit include?

The free audit returns a full scorecard: your AI-visibility score, a per-engine breakdown, share of voice against the competitors we detect, and the specific prompts you’re losing. To run it, enter your site and a work email at that same domain — we’ll email you a link, and confirming it signs you in (and verifies you as the domain’s owner, if no one’s beaten you to it) and lands you on the page where you click to run it. No card, ever. Each domain gets one free scan this way; re-scanning it later needs a paid plan.

How long does an audit take?

Most finish in 3–5 minutes; a slow one can take 15. You get live progress as each engine responds, and you can close the tab — the page keeps the outcome, and we email your report when there is one.

What exactly is in the fix pack?

The fix pack is generated from your own losing prompts: an FAQ page with FAQPage JSON-LD already embedded, a “you vs [rival]” comparison draft for each of your top rivals, a listings checklist of 10-15 places to get listed with a reason for each, and a generated llms.txt to serve at your domain root. Everything downloads as markdown you can publish as-is. During your trial the FAQ and your top comparison page download; the rest unlock with your first payment, or the moment you unlock early. After that the pack is regenerated after your weekly scan when your losing prompts change.

What counts as one "product"?

One product is one website: a single domain and the brand associated with it. Subdomains and additional landing pages on that same domain count as part of that one product, not separate ones. Owning a second, distinct website as its own product — its own fix pack, its own weekly scan — still needs a plan built for more than one, like Growth. Watching a second site you don’t own is different: private competitor watches are included on every paid plan, so keeping an eye on a rival doesn’t require upgrading.

Which AI engines are covered?

ChatGPT, Claude, Perplexity, and Google AI Overviews — the three surfaces buyers actually use today. We’ll add more once there’s a reliable way to measure them as honestly as these three.

How often do you re-scan?

Weekly, on paid plans. That’s the honest cadence at this price — trends need consistent sampling more than they need daily noise. Subscribers can also re-scan a product on demand from its page, once every 24 hours; a manual re-scan counts as that week’s scan, so the Monday scan (and its email or Slack digest) skips a product you’ve already scanned in the previous 5 days.

Can I change the competitors?

Yes. Signed-in users can edit their rival list at any time, and we re-score your already-stored answers against the new list instantly — no new scan required.

Is my scorecard public?

It lives at an unguessable link that’s excluded from search engines, so it won’t show up in a random search. Anyone you share that link with, though, can view it.

How does the trial work?

7 days, card required to start. Cancel anytime from the billing portal. Your first charge happens when the trial ends, not before. Your board seat goes up the minute the trial starts, labeled as a trial seat, and comes down if you cancel before that first charge. You can also end the trial early by paying — the charge happens that day and your billing month starts then.

Refunds?

We don’t refund partial months. If you think you were billed in error, email us and we’ll fix it.

How is the score computed?

Your score comes from three layers: which mention tier an answer gives you, where you land relative to other tracked brands, and a weighted mean across the engines that produced enough valid answers to qualify. Being an actual cited source counts for more than a link in the answer body, which counts for more than a bare name-check — that ordering is the whole idea behind the tiers. See the Methodology page for the exact formula and numbers.

Read the full methodology →

Who’s behind this?

Answergist is a self-serve AI-visibility audit and fix engine for founders who do their own marketing, showing you how ChatGPT, Claude, Perplexity, and Google AI Overviews describe your product. It’s operated by Wilphia LLC — hello@answergist.com. More on the About page.

More about Answergist →

Can my subscription sponsor my site automatically?

Yes. After you accept the Terms on a placement, paid invoices add to that listing’s all-time total on the front-page board — nothing extra is charged; it uses the subscription you already pay. A trial places a labeled seat at plan price: it sits below the paid rows and does not rank. Cancel before the first charge and the seat comes down. Paid invoices stay on the board even if you cancel later. Paying to sponsor someone else’s site does not steer your invoices there. It never changes your score. Only a refunded or disputed payment comes off the board.

How often is my Visibility Index listing rescanned?

The public snapshot updates on its own schedule — about once a month per listed site, via a daily cron that works through the roster in slices. Re-scanning your own dashboard (subscribers can do this on demand) refreshes your private report only; the public listing waits for its next scheduled pass, except for your very first scan right after you verify, which seeds it immediately.

Can I verify a site if I only have a Gmail (or other shared-provider) address?

No. Verifying ownership requires an email at your own site’s domain, like you@yoursite.com. Addresses at large shared providers — Gmail, Yahoo, iCloud, and similar — can’t prove you control a domain, so we don’t accept them. A shared inbox at your own domain, like hello@yoursite.com, works fine; if that leads to a dispute over who controls the account, contact us.

My product was already on the index — why is the score on my own dashboard different from the public listing?

A number of well-known products are seeded on the index before their owners ever verify, using competitors we detected automatically when we first scanned them. Once you verify one of those domains, your own scan detects competitors freshly from your site at that moment, which can differ from the set behind the public number. Your dashboard is the live one to act on; the public listing catches up on its own schedule, not from your dashboard scans.

Can I watch a competitor I don’t own?

Yes. Subscribers can add a private watch on any site — a competitor, anyone you want visibility into — and we scan it privately for you every week. Results live only on your own dashboard: we never publish a watch, list it anywhere, or change the watched site’s own public Visibility Index status. A watch never gives you access to that site’s own account or dashboard either — you only ever see the scans we run for you, never their private data.

How many prompts does it take to measure AI visibility reliably?

More than one, and the number matters more than most tools admit. Ask an assistant the same buying question on two different days and you’ll often get two different lists, so a single check tells you almost nothing. We run 15 buyer-intent prompts across four engines, which means a score is the mean of up to 60 answers rather than one lucky or unlucky reply. The prompt set is also frozen per version: if prompts were regenerated on every run, a week-over-week trend would be measuring prompt drift rather than your visibility. When we change how scoring works, that shows up as a version change, never as a silent jump in your history.

Why do ChatGPT and Google AI Overviews disagree about my brand?

Because they retrieve from different places — and the gap between them is usually the most useful number you have. Appearing consistently on one engine and being absent from another in the same category is common, and it normally tracks which sources each engine leans on rather than anything you changed. That’s why we publish a per-engine score beside the overall one: a 60 that’s 90/10/80 is a completely different situation from a 60 that’s 60/60/60. An engine also needs at least 8 valid answers out of 15 before it counts toward your overall score at all, and when one drops out we renormalize the remaining weights instead of scoring it zero, so an outage never reads as a collapse.

How can I check whether ChatGPT recommends my product without paying for anything?

Open ChatGPT, Claude and Perplexity, and run a signed-out Google search. Ask what a buyer would actually type — “best [category] for [use case]”, “[rival] alternatives” — five or six times over a couple of days, because the answers move. Write down who gets named and who gets cited. That is the manual version of what we automate, and it costs nothing. What a tool buys you is sample size, the same prompts every week, and a record you can compare against later. If all you want is a one-time gut check, do it by hand.

Is a citation worth anything if the AI never names my brand?

It is worth something, but less than being named, and we score the two differently rather than collapsing both into “mentioned.” Each answer counts at the highest tier it qualifies for: cited in the answer’s sources scores 1.0, a link in the answer body 0.75, your brand named in passing 0.5, and absent 0. Position then adjusts it — the first brand mentioned scores ×1.0, the second or third ×0.8, fourth or lower ×0.6. So one of your pages being cited while the answer recommends a rival by name is real, measurable, and still a losing outcome, which is exactly the gap the scorecard exists to show you.

Still stuck? hello@answergist.com.