← Back to Newsroom

AI Visibility Case Study: How We Used PR TIMES and GEO to Get Cited by AI

Big-O's own AI Visibility playbook — from PR TIMES syndication and trilingual Newsroom build-out to schema injection and AI Citation tracking.

AI Visibility Case Study: How We Used PR TIMES and GEO to Get Cited by AI
Case StudyBig-O Newsroom

1. Why PR TIMES as the Starting Point

When we decided to launch our own AI Visibility initiative, the first question was where to issue the first press release. We evaluated Taiwanese local outlets, English wires (PR Newswire, EIN Presswire), and Japan's PR TIMES. We picked PR TIMES for three reasons:

  • Multi-outlet syndication. A PR TIMES release auto-syndicates to Yahoo! JAPAN, Infoseek News, @nifty news, and 10+ other major Japanese outlets. For AI, this means "the same brand fact appears in many authoritative sources" — exactly the signal AI engines weigh for trust.
  • Japan-first market strategy. Our PCT patent filings prioritize Japan, so an early Japanese-language AI footprint is necessary market prep.
  • Cross-border AI training data. Japanese-language outlets are heavily ingested by ChatGPT, Gemini, Perplexity, and others. Publishing in Japanese plants the brand into multilingual AI knowledge bases.

2. Why GEO (Generative Engine Optimization)

Traditional SEO aims to rank #1 on Google. GEO aims for "when someone asks ChatGPT, Gemini, or Perplexity a question in your domain, the AI cites your brand." The optimization logic is fundamentally different:

AxisTraditional SEOGEO (Generative Engine Optimization)
Core signalsBacklinks, keyword densityStructured data, third-party verification, multi-source consistency
Content formLong-tail keyword articlesExplicit Q&A pairs, summarizable paragraphs, AI Summary blocks
Authority sourcesHigh-DA backlinksNews wires, Wikipedia, government open data
Tracking metricsRank, CTR, trafficAI citation count, AI citation ratio, brand-entity consistency

For B2B AI SaaS, GEO matters far more than traditional SEO — decision-makers increasingly ask ChatGPT "what AI sustainability solutions are out there?" instead of Googling.

3. A Three-Layer AI Search Strategy

  1. Content Layer. Every public article ships with an AI Summary at the top, explicit section headings, and FAQ at the bottom. AI crawlers extract the core message without parsing the full body.
  2. Schema Layer. Every article injects three JSON-LD schemas: Article (article metadata), Organization (brand entity reinforcement), and FAQPage (so AI can lift Q&A pairs directly).
  3. Cross-Language Layer. Three languages live on path-prefix routes (/zh-tw/, /en/, /ja/) wired together with <link rel="alternate" hreflang="...">. AI can find the language-matching citation when users ask in different languages.

4. AI Citation Design

AI Citation = an AI engine spontaneously names your brand when answering a user. To earn citations, three things must align:

  • Entity consistency. Brand name, founder, address, product lines must be identical across every page (own site, press release, Wikipedia, third-party directories). We use the exact same name, founder, address fields in the Organization JSON-LD everywhere.
  • Multi-source citation. The same fact must appear in multiple independent sources. The PR TIMES → Yahoo JP → Infoseek syndication chain creates 3+ authoritative outlets all mentioning "Big-O launched DSCS."
  • Structured evidence. Key claims (e.g., "world's first integrated framework," "Utility Model patent granted") carry dates, clause numbers, and externally verifiable links. AI lifts structured fields, not free text.

5. The Structured Data Checklist We Actually Ship

  • Organization — once at root layout. Includes name, legalName, foundingDate, founder, address, and sameAs (pointing to PR TIMES, LinkedIn, etc. as external verification anchors).
  • Article — per Newsroom post and Case Study. headline, datePublished, author, publisher, inLanguage, isPartOf.
  • FAQPage — the FAQ block on each article. This is where AI Citation hits most often.
  • BreadcrumbList — root layout. Helps AI parse site hierarchy.
  • NewsArticle (planned) — a more specific subtype of Article that Google News and AI Overview weight more heavily.

6. Schema Implementation Gotchas

  1. Canonical always points to the current language version, never the default. AI treats the canonical as the primary source.
  2. hreflang must be bidirectional. zh-tw → en + ja, en → zh-tw + ja, ja → zh-tw + en. Google Search Console flags unidirectional hreflang as an error.
  3. inLanguage in JSON-LD is mandatory, BCP 47 (zh-TW, en, ja). AI uses this to attribute the article to a language.
  4. FAQPage schema must match the visible Q&A on the page exactly — otherwise Google treats it as "hidden content" and disables rich snippets.
  5. Organization sameAs must be publicly reachable URLs. We list PR TIMES, LinkedIn, official GitHub.

7. Tracking & Validation

AI Visibility is a long-term investment, but short-term signals you can track:

  • Periodically prompt ChatGPT, Gemini, Perplexity, Claude with "what Taiwanese AI sustainability solutions exist?" and record whether the brand appears.
  • Monitor Google AI Overview for citation of our press release on relevant queries.
  • Verify the PR TIMES syndication chain reaches Yahoo JP, Infoseek, etc.
  • Watch server-side logs for AI crawler hits (GPTBot, PerplexityBot, ClaudeBot) and their frequency.

Closing

This is our own playbook, publicly documented step by step so other SMEs can use it as a template. If you'd like to start your own AI Visibility initiative, reach out at partnership@bigomorfi.ai.

Frequently Asked Questions

How is AI Visibility different from traditional SEO?
SEO optimizes for Google search rank. AI Visibility (GEO) optimizes for whether ChatGPT, Gemini, Perplexity, and Claude cite your brand. The core signals shift from backlinks and keyword density to structured data, third-party verification, and multi-source consistency.
Why PR TIMES instead of PR Newswire or Taiwan press?
Three reasons: (1) PR TIMES auto-syndicates to Yahoo JP, Infoseek, and ten-plus major Japanese outlets, generating multi-source consistency; (2) our PCT patent filings target Japan first; (3) Japanese-language outlets are heavily ingested by AI models, planting the brand into multilingual AI knowledge bases.
Do I really need hreflang?
If you want AI engines to find the right language version when users ask in different languages, hreflang is required. We use path-prefix routing (/zh-tw/, /en/, /ja/) paired with bidirectional hreflang link tags so both AI and Google can map languages correctly.
Which JSON-LD schemas should I inject?
At minimum four: Organization (once site-wide), Article (per post), FAQPage (per FAQ), and BreadcrumbList (site-wide nav). AI Citation lifts Q&A pairs from FAQPage most often, so FAQ quality directly affects citation probability.