AI Search Audit Entertainment · $1.6bn brand
taylorswift.com

The world's most searched name.
Five minutes from Blocked. Not five minutes from Cited.

GPTBot disallowed in robots.txt blocks OpenAI's direct citation surface. Removing it takes the rung from Blocked to Emerging (3/18) in roughly 5 minutes. The unblock is necessary but not sufficient: content gaps (no 2026 tour, mixed-reception 2025 album, post-Eras quiet period) cap category-query lift even with all bots fully open. Sustained Cited requires more than a robots.txt edit.

64
Composite Score
Developing
49Search
65Design
86Brand
56Compete
65Growth
Why this brand sits in a corporate audit

Taylor Swift Productions Inc. is a $1.6bn commercial enterprise: merchandise, touring (Eras: $2.2bn gross), licensing, and direct-to-fan e-commerce on taylorswift.com. Forbes, Bloomberg and the Wall Street Journal cover the operation as a branded business, not a personality. We audit the commercial domain against the same AI Search methodology used for Costco, Samsung and Nike: how citable is the official .com versus the third-party sites (Wikipedia, Spotify, Ticketmaster, Genius) that AI engines currently surface for fan and buyer intent queries. This is commercial commentary on a publicly accessible web property, not a personal critique.

Root cause · Access control
A site that's deliberately closed to the engines that now decide visibility.

Unlike Costco UK, whose problem is structural absence, taylorswift.com fails at a different layer. The infrastructure is live, the brand equity is enormous, but the site actively blocks GPTBot in robots.txt and offers no entity markup for the bots that do reach it. The result: AI engines route every fan and buyer query to Wikipedia, Spotify, Ticketmaster and Genius instead of the official commercial property. A 5-minute robots.txt edit moves taylorswift.com from Blocked to Emerging on the F&T /18 methodology; the lift to Cited then requires content additions.

01
GPTBot disallowed in robots.txt
OpenAI's training and retrieval crawler is blocked outright. ChatGPT cannot index, summarise, or cite content from taylorswift.com directly, so it cites Wikipedia and Spotify by default.
02
No MusicArtist or Event schema
The world's largest touring artist has no entity markup on her own domain. AI engines have no machine-readable signal that this is the canonical source for tour dates, releases, or merchandise.
03
Branded-house with broken internal graph
Dead internal links and JS-rendered navigation prevent crawlers from reaching the commerce surface. Even unblocked engines cannot map the store, the tour, or the catalogue from the homepage.
AI Visibility Maturity Curve · F&T /18 methodology
Blocked
AI Share of Voice (current) · Google 0/6 · Bing 0/6 · Brave 0/6 · Composite 0/18 (forced under substrate-level override)
Edit robots.txt to remove the Disallow: / line under User-agent: GPTBot. Under F&T's /18 methodology, the GPTBot Disallow rule triggers the Blocked rung override regardless of structural readiness; all 18 cells force to Unlikely. The unblock is a zero-cost 5-minute fix that restores OpenAI's direct citation pipeline. Post-unblock structural readiness is documented separately below.
Blocked
Emerging
Post-unblock projection (Step 4b)
5 minutes from Blocked to Emerging (3/18)
Projection lifts the Blocked override and recomputes against actual page content. You jump from Blocked to Emerging, skipping Invisible. The structural readiness is modest because content gaps (no 2026 tour, mixed-reception 2025 album, post-Eras quiet period) cap category-query lift. Per-engine projection: Google 1/6 · Bing 2/6 (taylorswift.com surfaces #1 for branded queries already) · Brave 0/6 (forum-bias prefers Reddit + editorial). Sustained Cited requires content additions: tour dates surface, biographical depth, Event + Person JSON-LD schema.
Full audit available
The 18-cell grid (current + projection rows), Phase B post-unblock projection panel, Citation Host Map for music/entertainment, Vendor Bake-Off Readiness (Partial 1/3), and the named Top Action all live in the full report.
Read the full AI Search Report →
Cohort callout · May 2026

Of 5 globally-famous brands tested under F&T's AI Search methodology in May 2026, 0 reached Recommended · 4 Cited · 0 Emerging · 0 Invisible · 1 Blocked.

The pattern across the Cited cohort: AI knows these brands by name, not by category. Branded queries surface the brand's own URLs; category queries route to editorial reviewers, comparison sites, or Reddit. The 1 Blocked brand is structurally different: robots.txt closure, not content gap.