【AI Visibility Case Study】SEO 92, But Invisible to AI Search — 52/100

Case Overview

This case features a water treatment chemical manufacturer with an established presence in its regional market. After completing a full AI visibility audit, we uncovered a striking contradiction: the company's traditional SEO performance was genuinely impressive — PageSpeed's SEO subscore came in at 92/100, and technical fundamentals like Sitemap, Meta Descriptions, and OG Tags were all properly configured. Clearly, the team behind the website knew what they were doing.

Yet when we simulated the queries that real procurement decision-makers type into ChatGPT, Claude, and Gemini — searches like "water treatment chemical suppliers" or "recommended coagulant manufacturers" — the company was completely absent. Traditional search engines could find it. AI search engines barely recognized it existed.

The overall AI visibility score came in at 52/100 — a mid-range result that signals untapped potential. But without intervention, this gap will widen rapidly as AI-driven search continues to displace traditional discovery methods in B2B procurement.

Score Breakdown

AI visibility isn't a single metric — it's the product of three distinct dimensions, each of which must be evaluated independently. A weakness in any one dimension creates a ceiling on overall performance.

The most critical bottleneck is the AI mention rate of just 35/100. Even with a solid technical foundation, the major language models are not including this company in their responses when answering industry-related questions. The GEO technical score of 47/100 reveals the root cause: the website lacks the structured markup that AI crawlers need to correctly interpret and categorize its content.

AI Search Visibility Testing

We submitted queries to three leading AI platforms — ChatGPT, Claude, and Gemini — to simulate how actual procurement decision-makers search for suppliers. Each platform received four queries: two brand-name queries and two industry-context queries, for a total of 12 tests designed to produce a statistically meaningful AI visibility snapshot.

Claude

Both brand-name queries on Claude returned vague mentions (△) — the model appeared to have fragmentary data about the company but could not confirm its market positioning or product specialization. More critically, both industry-context queries (e.g., "which companies manufacture water treatment chemicals in this region?") returned no mention at all (✗). Even users who already know the brand name may walk away with unclear information — and those searching by product category won't encounter the company at all.

ChatGPT

ChatGPT showed an inconsistent split: the first brand query returned a positive mention (✓), suggesting the company appears in OpenAI's training data. However, the second brand query produced only a vague mention (△), revealing reliability issues. Both industry queries returned zero mentions (✗). This kind of inconsistency is a serious liability for B2B companies — when two buyers ask the same question and receive different answers, it becomes nearly impossible to build a predictable brand presence in AI-generated recommendations.

Gemini

Gemini was the most consistent platform for brand recognition, returning positive mentions (✓) on both brand queries — the strongest brand performance across all three platforms. But the moment the query shifted to industry context ("recommend a water treatment chemical supplier"), Gemini also returned no mention (✗).

This pattern was remarkably consistent across all three platforms: the company has some brand recognizability, but is completely absent from high-purchase-intent industry queries. That's precisely the category of search that matters most in B2B — and the single largest AI visibility gap identified in this audit.

Across all 12 queries, the company received 6 mentions (50% mention rate overall). But every one of those mentions came from brand queries. All 6 industry-context queries returned zero results — revealing a fundamental disconnect between the company's current content strategy and the logic that AI recommendation engines use to surface suppliers.

Competitive Landscape

When AI platforms were asked to recommend water treatment chemical suppliers, the names that surfaced were predominantly large international players: Kemira, Nalco Water (an Ecolab company), Solenis, and several regional chemical groups from Japan and South Korea. These companies share a few characteristics that explain their AI visibility advantage: they publish extensive technical white papers in multiple languages, they have citations in academic and engineering journals, and their websites feature complete Schema structured markup alongside detailed case study pages.

By contrast, domestic water treatment chemical manufacturers are significantly underrepresented in the training data used by major AI platforms. Rather than treating this as a disadvantage, forward-thinking local manufacturers should recognize it as an open window. The companies that complete GEO optimization first will be positioned to claim recommendation slots that currently default to international brands — particularly for buyers who prefer or require local sourcing, responsive support, or compliance with domestic regulations.

The competitive gap is real, but so is the opportunity. First-mover advantage in AI visibility for this sector is still available — and the window won't stay open indefinitely.

GEO Technical Audit Results

The GEO technical audit evaluates the infrastructure that determines whether AI crawlers can correctly read, interpret, and cite a website's content. Think of it as the foundation layer of AI visibility — without it, even excellent content may go unrecognized.

The audit shows 5 out of 8 items passing at the baseline technical level, producing a GEO technical score of 47/100. Three missing elements combine to create what we call an "AI comprehension triangle" of blockers:

• Missing Schema JSON-LD means AI crawlers cannot identify the company's product categories, industrial applications, or business classification — they're guessing.
• Missing H1 Tags leaves each page without a clear topical anchor, making semantic understanding of individual pages unreliable.
• Missing Canonical URLs creates duplicate content risk, scattering brand signals across multiple URL variants and introducing noise into how AI models aggregate information about the company.

Together, these three gaps directly explain why brand queries produce partial recognition while industry-context queries return nothing. The AI models simply don't have enough structured signal to confidently place this company in a specific product category or recommend it for a specific application.

Website Performance

Website performance affects how deeply and how frequently AI crawlers index a site's content. Faster, more technically efficient sites tend to receive more complete crawls — which in turn supports stronger AI visibility over time.

The Performance score of 68/100 indicates room for improvement in load speed and Core Web Vitals — particularly Largest Contentful Paint (LCP) and First Input Delay (FID), both of which can reduce crawl efficiency. The SEO subscore of 92/100 is the standout figure: it confirms that traditional on-page SEO fundamentals are well-executed, which partially explains why the company still receives some brand mentions across AI platforms despite the structural gaps.

HTTP/2 has not been enabled, which limits the server's ability to handle multiple parallel resource requests efficiently. Enabling it is a relatively straightforward infrastructure upgrade with meaningful downstream benefits for crawl speed and overall AI visibility.

Expert Diagnostic Recommendations

Based on the audit data, three structural issues stand out as the highest-leverage areas for improving this company's AI visibility. The following diagnostics are intended to orient strategic decision-making rather than serve as a complete implementation guide.

Diagnostic 1 — Missing Semantic Markup Prevents AI Product Identification

The simultaneous absence of Schema JSON-LD and H1 Tags means that AI language models, when indexing the company's website, cannot reliably determine what the company's core products are or which industrial use cases they serve. Water treatment chemicals span an enormous application range — from municipal wastewater to semiconductor-grade ultrapure water systems. Without structured markup that segments these applications clearly, AI models can only surface the company with vague associations rather than precise, context-matched recommendations. A complete GEO optimization program should include custom Schema architecture for each major product line and application category.

Diagnostic 2 — Content Strategy Is Misaligned With AI Recommendation Logic

The fact that all six industry-context queries returned zero mentions is not primarily a brand awareness problem — it's a content matching problem. When AI platforms respond to queries like "recommend a reliable water treatment chemical supplier," they prioritize sources with demonstrable technical authority: application guides, dosing calculations, case studies with quantified outcomes, and specification comparisons. The company's current content inventory and structure does not yet satisfy these matching criteria. Addressing this gap requires a deliberate content strategy built around the specific questions that procurement engineers and environmental compliance officers are asking AI systems.

Diagnostic 3 — Missing Canonical URLs Create Hidden Signal Dilution

In traditional SEO, missing canonical tags are a manageable issue. In the context of AI crawling and knowledge aggregation, they represent a more insidious risk: when multiple URL variants carry similar content, the brand signals associated with that content get distributed across those variants rather than consolidated. This reduces the strength of the company's topical authority in AI model training and inference. Fixing canonical URLs is a low-cost, high-return technical task that should be prioritized early in any GEO optimization roadmap.

AI Search Trends in the Water Treatment Chemical Industry

Water treatment chemicals occupy a particular niche in B2B procurement: long decision cycles, high technical complexity, and buyers who invest heavily in research before making contact with any supplier. The typical decision-maker in this space — a municipal water authority procurement manager, an industrial plant environmental compliance engineer, an EPC contractor materials specialist, or an environmental consulting firm's technical advisor — does not make supplier selections lightly.

What's changing rapidly is where that research happens. Increasingly, the technical evaluation phase that used to involve hours of Google searches and PDF downloads is being compressed into direct AI conversations. An environmental engineer evaluating coagulant options for a high-suspended-solids wastewater challenge might ask ChatGPT: "What are the dosing differences between PAC and PAM for high-TSS industrial wastewater, and which suppliers are considered reliable?" In the traditional search era, answering that question required stitching together multiple web pages. In the AI search era, a single response does it — and that response includes a supplier shortlist. If a company isn't on that shortlist, it has been eliminated from consideration before the buyer ever visits a website or sends an inquiry.

Regulatory pressure is amplifying this trend. As environmental discharge standards tighten and carbon reporting requirements expand, procurement teams are increasingly searching for suppliers who can demonstrate specific technical compliance capabilities — removal rates for particular heavy metals, zero liquid discharge solutions, treatment performance under variable load conditions. These technically precise queries are exactly where AI search excels, and exactly where manufacturers with deep, well-structured technical content will dominate recommendation results.

The AI visibility competitive landscape in the water treatment chemical sector remains underdeveloped. Most domestic manufacturers have not yet implemented GEO optimization, which means the window for establishing first-mover positioning is still open. Companies that build out structured AI visibility now are likely to enjoy a 12 to 18-month advantage before the field becomes more competitive — and when AI platforms update their training data to include more structured industry content, early movers will capture a disproportionate share of recommendation exposure.

Disclaimer

This article is based on anonymized audit data from an actual AI visibility assessment. All information that could identify the specific company has been removed. AI platform responses are non-deterministic — the same query submitted at different times may produce different results. Technical audit scores and performance metrics reflect a specific point-in-time snapshot and may change as websites are updated.

Find Out Where Your Business Stands in AI Search

If your company operates in water treatment chemicals, environmental equipment, industrial chemistry, or a related B2B sector, now is the time to establish your AI visibility foundation — before competitors do.

🔍 Use our free AI Visibility Self-Assessment Tool to get a preliminary score for your business in under 5 minutes.

📞 Or book a free results consultation with our team for an in-depth analysis tailored to your industry and target buyer scenarios.

Explore more B2B AI visibility case studies at the Joseph Intelligence Case Study Index.