Skip to content

What one free scan actually checks

Five prompts, one engine, one page of results, no signup. Here is everything that comes back, and everything it deliberately leaves out.

What it runs

You give it a domain. Nothing else, no account, no card. Beseen works out the questions someone would ask an assistant when they are shopping in your category, takes five of them, and asks one engine live. The answers come back fresh, not read off something stored last month.

While that runs, it also fetches your site the way an AI agent would and checks whether the pages are actually usable: whether the content is there in the HTML, whether the structure holds up, whether a machine can follow it without a browser doing the work.

What comes back

  • Five prompts, asked once, on one engine.
  • Who the assistant names in each answer. If a competitor is cited and you are not, that is the finding.
  • Whether your own site is readable by agents at all, which is the difference between losing on content and never being in the running.
  • A result page with its own URL, so you can send it to a colleague or a client without them signing up either.

That is the whole output. One page, four things on it, nothing held back behind an email form.

What it does not include

The free scan is a sample, and being straight about the edges is more useful than overselling it.

  • One engine, not three. Paid plans ask ChatGPT, Claude and Perplexity.
  • Five prompts, not twenty-five. Five tells you whether you show up. It does not map a category.
  • One run, once. Nothing is watched over time and nothing gets re-checked.
  • No rewrites and no drafts. The scan shows you where you are absent. It does not hand you the copy that fixes it.
  • Competitor discovery, the keyword pull and the crawl of your pages all belong to the first full run on an account.
  • Nothing is changed on your site. Beseen never publishes anything, on any plan, free or paid.

The full first run is a different job: it finds your competitors, pulls the keywords you already rank for, and builds a set of 20 to 30 prompts off the back of both. It costs real money per run and it is slow, which is why the anonymous version is five prompts. A demo nobody waits around for is worse than no demo.

How to read the result

Two useful outcomes, and both are worth knowing. Either the assistant names competitors and skips you, which tells you the pages that would win those answers either do not exist or do not answer directly enough. Or it names you, in which case the question becomes whether that holds next month when everyone else starts writing for it.

One scan cannot answer the second one. That needs the same prompts asked again, week after week, against the changes you actually made. How it works covers that loop end to end.

After the scan

The free account is the next step up and still costs nothing: one site, 10 prompts on one engine, checked weekly, plus a brief and a prompt you can paste into your own tool for every page you are missing. Standard is $39 a month and adds 25 prompts across all three engines, with the rewrites written for you.

None of that matters until you know whether assistants name you today. Run the free scan on your domain and find out first.