Skip to main content
案例分析

【AI Visibility Case Study】High SEO, Zero AI Presence — 51/100

Published on March 24, 2026

Case Overview

This case features a Taiwan-based health supplement OEM and ingredient R&D company with strong product development capabilities and an integrated raw material supply chain. When we conducted an AI visibility audit on the company, we uncovered a striking contradiction: the company's website scored an impressive 92 on SEO with a solid technical foundation — yet it was completely invisible across AI platforms when buyers searched for industry-relevant terms like "Taiwan health supplement OEM manufacturers." The company did not appear in any AI-generated recommendation lists.

This finding highlights a critical truth: strong traditional SEO does not automatically translate into AI search visibility. The overall audit score came in at 51/100, placing the company in the "moderate potential" tier — with clear structural room for improvement.

Score Breakdown

The AI visibility audit evaluated the company across three dimensions. The score gaps paint a clear picture of where the problems lie:

Evaluation DimensionScoreStatus
AI Brand Mention Rate35 / 100⚠️ Needs Improvement
GEO Technical Health67 / 100🔶 Moderate
Website Performance (PageSpeed)57 / 100🔶 Moderate
Overall Score51 / 100⚠️ Moderate Potential

The most critical bottleneck is the AI brand mention rate of just 35 points. While some platforms responded to direct brand name queries, the company was completely absent from all "industry procurement scenario" searches across the three major AI platforms. This means that when potential clients use AI tools to find an OEM partner, this company simply doesn't make the shortlist.

AI Search Visibility Test Results

Genuine AI visibility isn't just about brand awareness — it's about whether a company gets recommended when buyers are actively looking for solutions. We submitted queries to three major AI platforms (Claude, ChatGPT, and Gemini), running 4 queries per platform for a total of 12 tests. These covered both brand-name queries and industry procurement scenario queries. The company was mentioned in 6 out of 12 tests. Here's how each platform performed:

Claude

In both brand-name queries, Claude returned a "vague mention (△)" — meaning Claude is aware the brand exists, but its descriptions were non-specific and failed to communicate the company's core capabilities or differentiation. More critically, both industry scenario queries (such as asking for recommended Taiwan health supplement OEM manufacturers) returned "✗ Not Mentioned." This indicates that Claude's knowledge base has only a surface-level understanding of the company, making it unable to proactively recommend the brand in procurement contexts.

ChatGPT

ChatGPT performed slightly better than Claude: the first brand query earned a "✓ Positive Mention," but the second dropped to "△ Vague Mention," revealing inconsistency. Both industry scenario queries returned "✗ Not Mentioned." This inconsistency reflects insufficient content density in AI training data — when query angles shift even slightly, recommendation outcomes fluctuate dramatically, making it impossible to establish a stable AI visibility position.

Gemini

Gemini delivered the strongest brand query results among the three platforms, returning "✓ Positive Mention" in both brand queries. However, even with this relative strength, both industry scenario queries still returned "✗ Not Mentioned." This result illustrates an important distinction: brand-level recognition and procurement-scenario visibility are two entirely different dimensions. The company has built some brand-level foundation, but has not yet established the content context needed for AI to prioritize recommending it when someone is searching for an OEM partner.

Across all three platforms, the 0/6 mention rate for industry scenario queries is the single most critical warning signal from this audit — it directly limits the company's ability to acquire new clients through AI-driven discovery.

Competitive Landscape Analysis

In industry scenario queries, the health supplement OEM companies that AI platforms recommended were predominantly those that have invested heavily in digital content. These competitors typically publish technical white papers, detailed OEM case studies, ingredient analysis articles, and have citation records in industry media and academic databases.

Among Taiwan-based competitors, several companies have built a clear "OEM expertise" identity within AI knowledge bases by consistently publishing structured content — GMP certification explanations, custom formulation case studies, and functional ingredient research summaries. By contrast, the company in this audit currently lacks the depth and structure in its public-facing content to allow AI to reliably identify its professional positioning across diverse query contexts. Closing this content gap represents the most actionable competitive opportunity available.

GEO Technical Health Check

GEO (Generative Engine Optimization) technical foundations determine whether AI crawlers can effectively parse and index a website's content. The audit examined 9 key technical indicators, with the company's website passing 8 of them — an 88.9% pass rate that exceeds the industry average.

Technical ItemStatus
Schema JSON-LD Structured Data✓ Configured
XML Sitemap✓ Configured
Meta Description✓ Configured
OG Tags (Open Graph)✓ Configured
Canonical URL✗ Not Configured
HTTP/2 Protocol✓ Enabled
Title Tag✓ Configured
H1 Tag✗ Not Configured

The two failing items carry significant consequences. A missing H1 tag means AI crawlers cannot quickly identify the core topic of each page, directly reducing content comprehension accuracy. A missing Canonical URL risks duplicate content issues that dilute AI attention across multiple page variants rather than concentrating authority on the primary page. Both fixes carry low implementation cost but produce notable GEO gains.

Website Performance

Website performance is one of the most overlooked factors in AI visibility — page load speed directly affects how deeply and how frequently AI crawlers index a site's content. The PageSpeed audit revealed the sharpest data contradiction in the entire report: an SEO score of 92 (excellent) versus a performance score of just 21 (critically low), yielding a combined PageSpeed score of 57/100.

This gap means the company's website performs well on traditional search engine metrics, but loads extremely slowly in practice. A performance score of 21 is the single most urgent fix identified in this audit. The likely culprits include uncompressed, oversized product images (common in health supplement photography), absence of CDN acceleration, and excessive JavaScript blocking page rendering. Slow page speed not only limits AI crawler indexing completeness — it also degrades the actual browsing experience for potential B2B buyers evaluating the company.

Expert Recommendations

Based on the audit data, we identified three optimization priorities with the highest return on investment:

1. Emergency Performance Fix: From 21 to 60+

A PageSpeed performance score of 21 is the single largest barrier to improving overall AI visibility. This score doesn't just hurt user experience — it actively constrains how efficiently AI crawlers can index the site's content. Diagnostics point to image asset handling and front-end resource loading strategy as the primary culprits, though a full technical roadmap requires deeper analysis of the site architecture to establish the correct remediation sequence.

2. Build a "Procurement Scenario" Content System

The root cause of a 0/6 industry query mention rate is that the company's existing content doesn't trigger AI recommendation logic in "find an OEM partner" contexts. Diagnostics point to a misalignment between content structure and Schema markup depth. While structured data is already configured, whether the markup is sufficiently detailed to support AI comprehension of OEM service scenarios requires a dedicated content audit to confirm.

3. Cross-Platform AI Visibility Distribution

The company's current performance varies significantly across AI platforms — Gemini is consistently positive, Claude is vague, and ChatGPT is inconsistent. This unevenness signals a structural gap in how the brand's digital content is distributed across the web. Identifying which external platforms and content formats can effectively close this gap requires a strategy tailored to the company's industry characteristics and competitive positioning.

AI Search Trends in Health Supplement OEM and Ingredient R&D

AI search tools are fundamentally reshaping how procurement decisions are made in the health supplement OEM industry — and the impact on small to mid-sized manufacturers is particularly significant.

Traditionally, health supplement brands found OEM partners through a familiar path: trade shows (such as Taiwan Food Show or CPHI) → peer referrals → direct search engine queries. Since 2024, however, a growing number of brand-side procurement teams have begun integrating AI conversation tools into their initial screening process. Typical queries now include: "Which Taiwan health supplement OEM factories hold NSF or USP certification?", "I want to develop probiotic gummies — which Taiwan OEMs have relevant formulation experience?", and "What's the difference between health supplement OEM and ODM, and how do I choose the right Taiwan supplier?"

These queries share a common profile: they are highly contextualized, carry specific intent, and frequently bypass brand names entirely to describe functional requirements. This means that unless an OEM manufacturer establishes clear associations within AI knowledge bases — linking specific functional ingredients, dosage form capabilities, and regulatory certifications together — even a reasonably well-known brand may fail to get recommended to the right buyers.

Ingredient R&D services introduce additional complexity: downstream clients may include health supplement brands, pharmacy retail chains, or overseas export buyers, each using different search languages and relying heavily on AI to provide comparative recommendations. Being named in an AI response as "a Taiwan company with R&D strengths in botanical extracts or microbial fermentation ingredients" represents a direct commercial opportunity.

Overall AI visibility in Taiwan's health supplement OEM industry remains low, with most companies yet to recognize the need for GEO optimization. This creates a first-mover advantage window. Companies that proactively build structured technical content, OEM case study libraries, and ingredient efficacy white papers in the coming months stand to gain AI recommendation advantages that will be difficult for competitors to overcome over the next 12 to 18 months.

Ready to See How Your Company Performs in AI Search?

The contradiction seen in this case — high SEO scores paired with near-zero AI visibility — is surprisingly common among Taiwan health supplement OEM companies. Is your company facing the same hidden competitive disadvantage?

Use our free AI Visibility Self-Assessment Tool to get a preliminary evaluation of your company in under 3 minutes. If you'd like to understand what the data means and how to prioritize your next steps, book a free results consultation with one of our advisors. For more AI visibility case studies and industry analysis, visit our Case Study Index.

Disclaimer

This article is based on anonymized real audit data with all identifying company information removed. AI platform responses are non-deterministic and may vary across different query sessions and time periods. Technical health check and performance scores represent a point-in-time snapshot.

FAQ

Why does a high SEO score not guarantee AI search visibility?
Traditional SEO optimizes for keyword ranking in search engines like Google by improving technical signals, backlinks, and on-page relevance. AI visibility, by contrast, depends on whether AI platforms like ChatGPT, Claude, and Gemini have enough structured, contextually rich content to recommend your company in response to real buyer queries. A technically clean website may rank well on Google but still be absent from AI-generated recommendations if it lacks the depth of content, industry associations, and semantic context that AI models rely on when generating answers.
What does an AI visibility audit measure that a standard SEO audit doesn't?
An AI visibility audit goes beyond technical SEO to assess three additional dimensions: how often and how accurately your brand is mentioned across major AI platforms (ChatGPT, Claude, Gemini), how well your site's structure supports generative engine crawling and comprehension (GEO technical health), and whether your content triggers recommendations in real procurement scenario queries — not just brand name searches. This provides a far more complete picture of your digital discoverability in an AI-first search environment.
How can health supplement OEM companies improve their AI visibility?
The most effective strategies involve creating structured, scenario-specific content that mirrors how buyers actually search for OEM partners. This includes publishing technical white papers on specific ingredients or dosage forms, documenting real formulation case studies, clearly communicating regulatory certifications (such as GMP, NSF, or ISO), and ensuring all of this content is properly marked up with Schema JSON-LD so AI crawlers can accurately interpret and categorize it. Building citation presence in industry media and trade publications also reinforces AI knowledge base associations.
What is GEO optimization and why does it matter for B2B manufacturers?
GEO stands for Generative Engine Optimization — the practice of structuring your website and content so that AI language models can efficiently crawl, understand, and cite it when generating responses. For B2B manufacturers, this is especially important because AI-driven procurement research is growing rapidly. Buyers increasingly use AI tools as a first-pass filter before contacting suppliers. If your company isn't surfaced in those AI responses, you're effectively invisible to a growing segment of high-intent prospects before they ever reach your website.
How does page speed affect AI visibility for B2B websites?
Page load speed impacts AI visibility in two ways. First, AI crawlers — like all web crawlers — have limited crawl budgets per site. Slow-loading pages reduce how deeply and how frequently crawlers can index your content, which means important pages may be missed or underweighted in AI training data. Second, slow page speed degrades the experience for human buyers who arrive via AI-recommended links, increasing bounce rates and reducing the behavioral signals that reinforce your credibility. A PageSpeed score of 21, as seen in this case, represents a serious obstacle to both AI indexing efficiency and buyer conversion.
How long does it take to see improvements in AI visibility after optimization?
Results vary depending on the scope of changes and how frequently AI platforms refresh their knowledge bases. Technical fixes like adding H1 tags, configuring Canonical URLs, and improving page speed can have relatively quick effects on crawler behavior — often within weeks. Content-driven improvements, such as building out structured case studies and ingredient white papers, typically take 3 to 6 months to meaningfully influence AI recommendation patterns. Companies that start GEO optimization now are likely to see compounding advantages within 12 to 18 months as AI search adoption in B2B procurement continues to accelerate.

Ready to Elevate Your Digital Marketing?

Let our AI-driven solutions help your business grow