Tutorials
Subscription Client Creation
Visibility Audit Scheduled Audits Dashboard
Client Portal Showcase Page White Label
MCP Guide REST API AI Tracking
Source Authority

Control the URLs AI cites about your brand

Source Authority extracts every URL AI models reach for when they answer questions about you — owned, earned, competitors, hallucinated. Read your sources, plug the leaks, take ownership of your narrative.

8 min Consultant

Table of Contents

  1. What is Source Authority?
  2. Why it matters
  3. How it works under the hood
  4. Reading the dashboard
  5. Hallucinated URL alert
  6. Typical use case
  7. Plan and access
  8. Going further

What is Source Authority?

Source Authority is the AI Labs Audit feature that extracts the URLs actually cited by AI models when they answer questions about your brand. For each prompt run during an audit, every web link surfaced by ChatGPT, Gemini, Perplexity, Claude, Mistral or Copilot is parsed, categorised and aggregated.

You end up with three actionable lists:

  • Owned — URLs that belong to your domain and your sub-brands.
  • Earned — third-party URLs reinforcing your visibility (press, directories, Wikipedia, partner sites, your competitors).
  • Hallucinated — URLs that look plausible but do not actually exist on the open web.

Why it matters

Traditional ranking told you which page comes up first on a search results page. AI ranking is different: the model decides which sources to quote when it composes its answer. If your competitor’s Wikipedia entry is the model’s favourite, your name still appears, but the credibility lives on someone else’s URL.

  • Spot the pages that build your AI credibility so you can keep them up to date.
  • See which competitors are capturing your mentions on their territory (their blog, their comparator, their directory listing).
  • Catch hallucinated URLs before a prospect clicks on a fake link and lands nowhere.
  • Build a content roadmap based on real AI behaviour, not on a guess about what “the algorithm” might prefer.

How it works under the hood

The pipeline is automatic, no extra setup is required:

  1. Audit run. The audit fires a set of branded and unbranded prompts against the AI models you selected.
  2. URL extraction. Each model response is parsed: links, citations, footnote-style sources, in-text URLs — everything is captured, even when the model wraps the URL in markdown or HTML.
  3. Classification. Each URL is matched against your client domain (and configured aliases) to flag owned, against your declared competitors to flag competitor, and otherwise grouped under earned.
  4. Reachability check. A lightweight HTTP call confirms the URL exists. URLs that return 404 / DNS error / connection refused are tagged hallucinated.
  5. Aggregation. Citations are counted per domain, per model, per market and per audit so you can compare frequency over time.

Reading the dashboard

In the Dashboard v4, Source Authority lives in Zone 2 — Presence in AI answers, in a panel called “Cited sources”. You see a horizontal bar chart of the top-cited domains, ordered by frequency.

You can switch the panel between three views:

  • Per domain — the default. Shows where AI sends users.
  • Per model — useful when ChatGPT and Perplexity diverge on which sources they favour.
  • Per market — for multilingual audits, see whether Wikipedia FR cites you while Wikipedia EN does not.

Hallucinated URL alert

When an AI model fabricates a URL — typically a path on your own domain that does not exist — the row gets a red “non-existent URL” badge and the alert is repeated in Zone 5 — Action plan with a “fix or redirect” recommendation.

You have two responses:

  • Create the missing page if the model invented a path that would actually be useful (FAQ, pricing, comparison) — you turn an hallucination into real organic traffic.
  • Set up a 301 redirect from the hallucinated path to the closest real page so that any user clicking that link still lands on you.

Typical use case

A B2B SaaS audits its visibility on prompts like “best CRM for SMBs”. Source Authority shows that 62% of the URLs cited come from a Wikipedia entry maintained by a competitor and a third-party listicle that ranks the competitor first. Their own domain only accounts for 11%.

The reaction plan, generated automatically in Zone 5:

  • Update the Wikipedia entry with neutral, sourced facts about the brand.
  • Publish a comparison page on their own domain, with structured data, to become the canonical source.
  • Reach out to the listicle author for a re-evaluation pass with new evidence.
  • Track the next audit: target ratio above 35% owned within two months.

Plan and access

Source Authority is included from the Consultant plan upwards. The Consultant+, Agent and Agence+ plans add the per-model and per-market views. The Discovery plan only displays the global domain ranking, without the hallucination check.

Ready to audit your AI visibility?

Create your free account and receive 600 bonus credits.

Create free account