Skip to main content

How it works

From one URL to a plan you can act on

Rankin AI reads your public pages the way a crawler would, grades them against a fixed rubric, and turns each gap into a concrete fix.

  1. 1

    Tell us about the site

    Give us the URL, your business name, industry, audience, main services and location. Add up to five competitor URLs if you want a comparison. You confirm you are allowed to audit the site.

    • We vet the address before a single request goes out. Only http and https, no credentials in the URL, and no private, reserved or cloud-metadata addresses.
    • The business details feed one specific check: do the words you use for your services actually appear on your pages?
  2. 2

    We crawl it

    A background worker grabs your robots.txt, sitemap and llms.txt, then reads up to 50 public pages, depending on your plan. The pages that matter most for AI visibility readiness go first.

    • It introduces itself as RankinAI-Auditor and obeys any robots.txt rules aimed at that name.
    • No forms submitted. No logins. No downloaded files run. Cart, checkout, account and login pages are skipped outright.
    • Homepage, about, services, FAQ, contact, case studies and articles come before deep or parameter-heavy URLs.
  3. 3

    We score seven categories

    About seventy checks run on what the crawler collected. Each earns some share of its points and writes down exactly what it found.

    • Scoring is deterministic. Feed it the same pages and you get the same number back. Nothing is random.
    • If a check does not apply to your site, it drops out of the math entirely, so you are never marked down for something irrelevant.
    • An optional AI pass can smooth the wording of summaries. It cannot touch a score or rewrite evidence.
  4. 4

    You get a ranked to-do list

    Every check that fails or only partly passes turns into a task. Tasks are sorted by severity, then impact, then effort, and split into three time horizons.

    • Do first: high-impact fixes, most of them doable in one sitting.
    • Next 30 days: bigger improvements that need some coordination or writing.
    • Longer term: slower bets that compound, like original research and in-depth service pages.
  5. 5

    You run it again

    Make your fixes, then check again. The report shows how your score moved since the last run on that domain.

    • Score history comes with Growth and Agency plans.
    • Since scoring is repeatable, a move in the number means something changed on your site, not in the measurement.

Methodology

How your overall score is built

Every category gets its own 0-100 score from its own checks. The overall score is a weighted blend of the categories we could measure.

Category score

category_score = round(100 × Σ(points_awarded) / Σ(points_possible))

  — summed only over checks that apply to this website
  — a check that cannot apply is excluded from both sums

Overall score

overall = round( Σ category_score × effective_weight )

effective_weight = base_weight / Σ(base_weight of available categories)
Category base weights
CategoryBase weight
Technical AccessibilityWhether AI crawlers and answer engines can reach, load and index your pages at all.15%
Entity ClarityHow unambiguously your site states who the business is, what it does and who it serves.17%
Content AuthorityDepth, originality and expertise signals in the content a model would draw on.18%
Answer ReadinessHow easily individual passages can be extracted as direct answers to real questions.17%
Structured DataMachine-readable schema markup describing your organization, content and offerings.13%
Trust and EvidenceAuthorship, contact details, citations and proof that support what the site claims.15%
Competitive VisibilityHow your observable signals compare with the competitors you supplied.5%

When there are no competitors to compare

Skip the competitor URLs, or give us ones we cannot reach, and Competitive Visibility is marked unavailable. Its 5% weight is spread across the other six categories in proportion to their own weights. Your overall score stays on the same 0-100 scale.

The worked examples, the complete check list and every point allocation live in AUDIT_METHODOLOGY.md in the product repository.

Straight talk

What this check can and cannot tell you

Rankin AI measures signals that are visible on your website. That is useful. It is not a crystal ball, and we want to be exact about the difference.

What you get

  • Signals read straight from your site, with the evidence attached
  • An AI visibility readiness score from a fixed, published rubric
  • A short list of changes that may improve discoverability
  • A side-by-side against the competitors you name
  • Optional public search results, labeled as exactly that

What nobody can promise

  • That ChatGPT, Perplexity, Gemini or Copilot will name your brand
  • A look inside any AI system’s training data, index or ranking logic
  • A position, a citation rate or a number of referrals
  • A count of how often a private AI product mentions you today
  • A substitute for publishing something genuinely useful

Every report keeps four things apart: signals measured on your site, optional public-web search observations, AI-written interpretation, and sample or demo data.

Browse every check we run

See it work before you spend a dollar

The instant demo takes about five seconds. The free preview runs on your real homepage, no account required.

Try the free demoSee pricing

Pay once from $49 or subscribe from $29 a month. Every price is in US dollars.