If none of the above matches what you are seeing, it is worth separating two different kinds of problem before you go looking for help, because they have different routes. If the question is how to use something, open Get Help at the bottom of the sidebar. It carries Guided Tours that walk through the screens step by step, which is usually faster than describing a workflow in writing and waiting on a reply. If something looks wrong rather than unfamiliar, start with Troubleshooting, which covers the symptoms people report most often and what usually turns out to be behind them. Most reports resolve to one of three causes and are answered there faster than by waiting on a reply. If that does not cover it, get in touch through seerly.app/contact, or write to hello@seerly.app if you would rather use email. Either reaches the same place. Whichever you use, a handful of details will save a round trip. Include: - The domain and project you are looking at, since many accounts have several and the answer frequently depends on which one. - The exact screen, written as the path you took to reach it, for example Sidebar → Insights → Visibility. Several screens show similar-looking numbers, and knowing which one you were on removes most of the ambiguity. - What you expected against what you saw. This sounds obvious and is the single most useful thing you can write, because a number that looks wrong is often a number being read against a different definition than the one it uses. - When you noticed it, roughly, and whether it has been consistent since. A problem that appeared once and cleared is a different investigation from one that has persisted for a week. - How wide it goes: one topic, one engine, or everything. That single detail usually determines where the answer lies, because a problem confined to one engine points somewhere quite different from one affecting the whole account. - A screenshot if the issue is visual, cropped to the panel in question. If your question is about billing, plans or what your subscription includes, say so up front and skip the rest, since none of the diagnostic detail applies and it will only slow the reply down.

Frequently asked questions

Q: Which AI engines does Seerly track?

Seerly tracks ChatGPT, Perplexity and Google AI Overview. The standard plans, compared at seerly.app/pricing, cover combinations of those three, and your own coverage is shown in Sidebar → Settings. Claude and Gemini are available on enterprise arrangements, agreed as part of a custom plan rather than picked from the pricing page. Engines disagree with each other far more often than people expect, because they draw on different sources to build their answers. That makes the per-engine breakdown considerably more useful than the overall average, which can hide a total absence on one surface behind a strong showing on another. See AI engines tracked.

Short answers to the questions we get most often about how Seerly works.

HelpUpdated 25th August 20266 min read

How is Seerly different from an SEO tool?#

SEO tools measure where you rank in a list of links. Seerly measures what AI assistants actually say when somebody asks for a recommendation in your category: whether you get named at all, how you are described when you do, and which sources the answer leaned on to reach that description.

The two are not substitutes. A brand can rank first on Google and go entirely unmentioned by every AI assistant, because the assistant is composing an answer from sources it trusts rather than ordering a list of pages. Most companies discover this the first time they ask an assistant about their own category.

Which AI engines does Seerly track?#

Seerly tracks ChatGPT, Perplexity and Google AI Overview. The standard plans, compared at seerly.app/pricing, cover combinations of those three, and your own coverage is shown in Settings. Claude and Gemini are available on enterprise arrangements, agreed as part of a custom plan rather than picked from the pricing page.

Engines disagree with each other far more often than people expect, because they draw on different sources to build their answers. That makes the per-engine breakdown considerably more useful than the overall average, which can hide a total absence on one surface behind a strong showing on another. See AI engines tracked.

Do I need to install anything on my website?#

No. There is no code, no tracking pixel, no DNS change and no plugin, which also means nothing for your developers to schedule. Seerly analyses your public site and public AI answers entirely from the outside, the same way an engine would. You enter a domain and it starts.

Connecting analytics is optional and only worth doing if you want to see downstream traffic alongside your visibility numbers. See Integrations.

How long does the first analysis take?#

Typically about an hour, and up to 4 to 6 hours on Pro, since wider coverage means more engines and more prompts and there is simply more to run.

Progress is shown live, stage by stage, so you can watch it complete rather than wonder. You do not need to stay on the page: Seerly emails you when the dashboard is ready, and closing the tab does not pause anything. See Getting started.

Why haven't my numbers changed since yesterday?#

Because your analysis has not re-run yet. Numbers here move when a run completes rather than continuously, so a day with no run is a day with no change by design.

Check when the last run finished on InsightsVisibility. Your plan sets how often runs happen; see seerly.app/pricing.

Why do the results change between runs when nothing changed?#

Because AI answers are generated fresh every time and vary naturally as a result. This is expected behaviour rather than a fault, and it is precisely why Seerly asks a fixed set of prompts repeatedly instead of asking once and reporting whatever came back.

The practical consequence is that any single run is one data point rather than a finding. Judge trends across several runs, which is what Reading trends covers in detail.

Can I track competitors?#

Yes, on every plan. Seerly identifies likely competitors from your category and you manage the tracked list on InsightsCompetitors using Select Competitors.

Tracked competitors are measured on exactly the same prompts as you, which is what makes the comparison fair rather than two separate reports placed side by side. It is worth reviewing that list early, since every comparison in the product is drawn against it. See Competitors.

Can I manage several brands in one account?#

Managing several brands from one account depends on your plan. See seerly.app/pricing.

If a domain already belongs to a project under somebody else's account at your company, you will be offered the option of requesting access to that project rather than creating a second one, and it is worth taking. Two disconnected projects covering one brand split the history between them and produce two sets of numbers that disagree.

Why don't the tracked topics match what we sell?#

This is the most common thing people report, and it is worth fixing before anything else, on Fix it on ActionsPromptsManage. Everything in the product is measured against your topics, so a topic set that has drifted makes every downstream number an accurate description of somebody else's market.

Check the target country while you are there, since it scopes the whole picture to a single market and a wrong one produces a confident, coherent view of somewhere you do not sell. See Topics and prompts.

Why are we invisible on one engine but fine on another?#

This is common, and it is usually a source problem rather than anything to do with your brand. The engine you are missing from is drawing on sources that do not mention you, while the engine where you do well is drawing on ones that do.

Look at what that engine cites instead for the topic in question, because that is the gap worth closing. More often than not it turns out to be a review site or a directory rather than a competitor, and getting listed is faster than publishing. See Gap analysis.

Why does nothing we publish get cited?#

There are two things to check, and the order matters.

First, whether AI crawlers can read your pages at all. Smart Audit covers this, and it is worth confirming rather than assuming, because a site can look perfectly healthy to Google while serving almost nothing to everything else, which happens particularly often when pages are assembled in the browser.

Second, whether your pages answer questions specifically enough to be quoted. Engines lift direct, concrete claims out of pages, and general marketing prose gives them nothing to work with. Allow some time as well: publishing, discovery, indexing and citation span several cycles between them.

Still stuck?#

If none of the above matches what you are seeing, it is worth separating two different kinds of problem before you go looking for help, because they have different routes.

If the question is how to use something, open Get Help at the bottom of the sidebar. It carries Guided Tours that walk through the screens step by step, which is usually faster than describing a workflow in writing and waiting on a reply.

If something looks wrong rather than unfamiliar, start with Troubleshooting, which covers the symptoms people report most often and what usually turns out to be behind them. Most reports resolve to one of three causes and are answered there faster than by waiting on a reply.

If that does not cover it, get in touch through seerly.app/contact, or write to [email protected] if you would rather use email. Either reaches the same place.

Whichever you use, a handful of details will save a round trip. Include:

The domain and project you are looking at, since many accounts have several and the answer frequently depends on which one.
The exact screen, written as the path you took to reach it, for example InsightsVisibility. Several screens show similar-looking numbers, and knowing which one you were on removes most of the ambiguity.
What you expected against what you saw. This sounds obvious and is the single most useful thing you can write, because a number that looks wrong is often a number being read against a different definition than the one it uses.
When you noticed it, roughly, and whether it has been consistent since. A problem that appeared once and cleared is a different investigation from one that has persisted for a week.
How wide it goes: one topic, one engine, or everything. That single detail usually determines where the answer lies, because a problem confined to one engine points somewhere quite different from one affecting the whole account.
A screenshot if the issue is visual, cropped to the panel in question.

If your question is about billing, plans or what your subscription includes, say so up front and skip the rest, since none of the diagnostic detail applies and it will only slow the reply down.