Case studies provide concrete evidence for how strategies perform in the real world. In the rapidly shifting landscape of search, where large language models (LLMs) and AI assistants now assemble answers from multiple sources instead of serving a list of blue links, traditional search engine optimisation (SEO) metrics no longer tell the whole story. A brand can retain high rankings in Google’s organic results yet disappear entirely from AI‑generated responses. Being recognised within these generative systems depends on entity clarity, structured data, and content that clearly answers questions. Businesses must therefore look beyond click‑through rates to understand how they are represented in AI answers. This case study examines how Chemours, one of the world’s largest producers of titanium dioxide (TiO₂), transformed its presence in AI search results by embracing generative engine optimisation (GEO) and answer engine optimisation (AEO). It shows that with the right tactics, even a technically complex B2B brand can go from obscurity in AI assistants to dominant share of voice.

Background: The Initial Challenge

Chemours operates multiple facilities across North America and supplies titanium dioxide pigments used in coatings, plastics, and other industrial applications. Despite its size and established authority in traditional search results, the company faced a fragmented digital presence. Each region maintained its own website and technical documentation, leading to inconsistent messaging, duplicate or conflicting content, and a patchwork of technical standards. While this structure didn’t prevent Chemours from ranking in conventional search, AI assistants such as ChatGPT and Perplexity struggled to interpret and cite the brand. As generative search experiences began to roll out, the marketing team noticed declining click‑through rates on high‑ranking pages and a rise in “zero‑click” behaviour, where users obtained the information they needed directly from AI summaries. Competitive brands started appearing in AI answers for key queries like “best titanium dioxide for plastics” or “titanium dioxide coatings manufacturer,” while Chemours received few or no mentions. Baseline analysis revealed that the brand was rarely cited in Google’s AI Overviews or conversational answers and had virtually no visibility across ChatGPT or Perplexity for commercial or informational queries. Although the company’s pages continued to rank highly in Google’s SERPs, it wasn’t the brand being recommended by AI systems.

Baseline Metrics Before GEO/AEO

Before undertaking the generative optimisation initiative, Chemours’ marketing analytics showed:

• Strong organic rankings for industry‑specific keywords related to titanium dioxide and coatings, with top‑three positions for many technical terms on Google.
• Flat or declining click‑through rates (CTR) despite stable rankings, suggesting that AI answers were fulfilling user intent without a click.
• Minimal brand mentions in AI answers across ChatGPT, Perplexity, and Bing Copilot; some queries produced generic definitions without referencing any manufacturer.
• No citations in Google AI Overviews for formulations or coating questions, even though Chemours published detailed product data.
• Weak performance on comparison and recommendation prompts, such as “best white pigment for plastics” or “titanium dioxide alternatives.” The brand rarely appeared in AI‑generated lists, indicating poor entity representation.

The company realised that its content, while comprehensive, was not structured in a way that AI systems could easily interpret or summarise. Its technical documents were long and lacked concise overviews, product specifications lived in PDFs with little structured metadata, and there was no unified schema across regional sites. Without clear signals to connect content to entities and topics, language models could not identify the brand as an authoritative source.

The GEO/AEO Strategy Implemented

To address these challenges, Chemours partnered with the ABM Agency, which specialises in generative optimisation. Together they devised a strategy built around answer‑first content and technical alignment across all digital properties.

From Keyword‑First to Answer‑First Content

Instead of optimising for specific keywords, the team focused on the questions potential customers were asking in AI assistants and voice interfaces. They mapped high‑intent queries across the customer journey, from top‑of‑funnel definitions (“What is TiO₂?”) to comparison queries (“rutile vs. anatase titanium dioxide”) and commercial prompts (“best titanium dioxide for coatings”). Each query was assessed for its likelihood to trigger AI responses in Google’s AI Overviews, Bing Copilot, ChatGPT or Perplexity. Queries that tended to elicit answer boxes or generative snapshots were prioritised, and content was designed to deliver clear, concise answers that could be extracted directly into AI summaries.

Unifying the Web Presence

One of the biggest technical challenges was the fragmented web infrastructure. Chemours had 12 regional websites using different content management systems (CMS) and inconsistent taxonomies. The ABM Agency consolidated these into a single platform with a common information architecture. Page templates were rebuilt to support structured data and fast load times, achieving sub‑two‑second page speed across all global markets. Consolidation also ensured consistent messaging and allowed the team to centralise product specifications into a comprehensive database. This unified environment made it easier to apply schema markup and maintain entity consistency across languages and regions.

Structured Data and Entity Clarity

The team implemented extensive schema markup, using types such as Product, Article, FAQ, HowTo, and BreadcrumbList. Product pages included aggregateRating, manufacturer, material, application, and other attributes. Where appropriate, they used PropertyValue entries to label pigment properties like particle size, brightness, and application categories. The addition of FAQPage and HowTo schema made it easier for AI systems to extract question‑answer pairs. Each product and category page opened with a succinct summary that defined the product, its main use cases and competitive advantages. By embedding definitions within the first paragraphs and using consistent terminology, the site signalled strong entity clarity. Internal linking was tightened to reinforce topical depth; pages referenced related products, processes, and application guides, ensuring that AI models could map relationships between topics.

Content Redesign for Quotable Summaries

Long technical documents and scattered PDF spec sheets were transformed into digestible web content. Each page was restructured around a problem‑solution framework: a brief introduction that defined the pigment or application, followed by numbered steps or bullet points outlining key features, benefits, and performance data. Executive summaries and TL;DR sections allowed AI engines to grab coherent snippets. Comparison pages were developed to answer prompts like “anatase vs. rutile TiO₂,” featuring tables that compared properties, cost considerations, and best‑fit applications. The team also created guides on choosing pigments for specific industries (plastics, coatings, paper) and wrote how‑to articles on formulating paint or plastics with TiO₂. These guides were broken into question‑and‑answer sections, making them natural candidates for extraction by generative models.

External Authority and Trust Signals

Beyond the website overhaul, ABM Agency knew that AI systems look for trusted third‑party sources when constructing answers. Chemours’ subject‑matter experts collaborated with industry journals and technical publications to publish papers and benchmarking studies. They produced original guides comparing chloride versus sulfate processes for TiO₂, citing research and data. These pieces were hosted on the company’s site and cross‑linked from reputable industry portals. The company also invested in thought‑leadership content, including webinars, whitepapers, and conference presentations, each with clear authorship and update dates. Author bylines were connected to professional profiles with verifiable credentials, satisfying the “experience, expertise, authority, and trust” (E‑E‑A‑T) signals that AI systems evaluate. By reinforcing their authority through external citations and consistent author attribution, they increased the likelihood of being referenced by AI models.

AI Visibility Testing and Monitoring

Throughout the implementation, the team regularly tested Chemours’ visibility across AI platforms. They issued prompts on ChatGPT, Perplexity, Bing Copilot and Google’s SGE for targeted queries and recorded whether the brand appeared in the answer, whether it was cited as a source, and how prominently it was mentioned. They logged metrics such as AI citation count, brand mention frequency, and the prominence of citations (main answer vs. supporting footnote). For each iteration of content updates, they compared results against previous baselines. This systematic testing helped identify which pages were successfully triggering AI citations and which needed further improvements. The team also tracked referral traffic coming from AI interfaces using unique UTM parameters and Google Analytics 4 event filters. Even though AI traffic was initially small, they noted that visits from chatbots showed longer engagement and higher conversion rates.

Results: What Changed

The generative optimisation program produced dramatic improvements within a few months. According to the SE Ranking report:

• ChatGPT mention rate reached 82% for coatings and plastics queries. Before the initiative, Chemours was hardly mentioned; afterwards, more than four out of five relevant prompts included the brand in ChatGPT answers.
• Google AI Overviews referenced Chemours in 84% of paint and coating formulation queries, signalling a strong presence in Google’s generative search experience.
• Perplexity citations rose to 73% in competitive evaluations of titanium dioxide, giving Chemours a significant share of voice against rivals.
• Competitive advantage: The unified strategy gave Chemours a mention share of 67% versus 23% for its rival Tronox and 18% for Venator in AI responses. This dominance made the brand the default choice in many AI answers.
• Business impact: The improvements influenced over $90 million in pipeline and more than $20 million in revenue, demonstrating that AI visibility translates into tangible commercial results. High‑intent leads came from conversations in AI assistants where buyers encountered Chemours as a recommended provider.

These results illustrate how generative optimisation goes beyond ranking and clicks; it positions a brand as the authoritative answer. Chemours moved from near invisibility in AI answers to a dominant presence, capturing the attention of early‑stage researchers and ready‑to‑buy customers alike. The brand’s message became consistent and extractable, and AI systems began to view its pages as canonical sources for titanium dioxide topics.

Business Impact Beyond Clicks

The improvements were not just numbers on an analytics dashboard. Sales and marketing teams reported that prospects came into consultations with a better understanding of Chemours’ products and differentiation. Because AI responses already provided clear definitions and comparative advantages, conversations shortened and conversions accelerated. Sales calls often began with prospects saying they had seen Chemours referenced by AI and were interested in specific formulations. Customers arriving from AI referrals spent more time on site and viewed more pages per session, indicating higher intent. Internal customer experience surveys showed improved brand recall among buyers who had used AI tools for research, reflecting the trust built through consistent representation in generative responses.

Illustrative Real‑World Parallel

Chemours’ experience underscores a wider trend: brands that invest in authoritative, structured content become the default recommendations in AI answers. Consider HubSpot, a CRM and marketing software provider. When users ask ChatGPT or Perplexity “What’s the best CRM for small businesses?” or “How do I choose a marketing automation tool?”, HubSpot frequently appears among the top suggestions. This is not just because HubSpot ranks highly in traditional search results; it’s because the company publishes clear guides on CRM selection, comparison tables, and definitions of key concepts. Their educational content is structured with FAQs, how‑to sections, and schema markup that signals entity relationships. As a result, AI models identify HubSpot as a trusted source for CRM-related questions and cite it accordingly. The same principles that elevated Chemours—clarity, structure, and authority—explain why HubSpot is repeatedly recommended in AI answers.

What Didn’t Work (and Was Adjusted)

Not every tactic delivered immediate success. Early in the project, the team produced long, detailed articles on manufacturing processes without summarised introductions. AI assistants rarely extracted from these pages because the key information was buried several paragraphs down. They also added schema markup indiscriminately, marking up pages without providing concise answers or meaningful questions. This resulted in zero improvement in citation rate because the underlying content still lacked clear, quotable statements. The team learned that structured data is only effective when paired with answer‑oriented writing. Similarly, some pages that ranked well in Google were initially left unchanged under the assumption they would perform well in AI search. These pages continued to fail in AI citations until they were rewritten to emphasise definitions, comparisons, and facts. The biggest lesson was that AI visibility isn’t a by‑product of traditional SEO success; it requires explicit attention to how content is formatted and summarised.

Key Lessons Learned

1. AI visibility is cumulative: It takes consistent reinforcement across multiple pages and platforms to establish a brand within AI knowledge graphs. A single optimised article will not make a difference if the rest of the site remains unclear.
2. Structure and clarity outperform volume: Short, well‑structured passages and bullet points that address specific questions have a higher chance of being cited than lengthy treatises. Generative models prefer content that can be broken into chunks and summarised.
3. Entity authority matters more than keyword ranking: AI engines rely on semantic associations. Consistent terminology, accurate definitions and strong entity linking help models map a brand to a topic. Without entity clarity, high rankings don’t translate into citations.
4. Freshness and authority go hand in hand: Regularly updating content, citing new research, and providing current data signals to AI systems that the information is trustworthy and relevant. Outdated pages are less likely to be surfaced.
5. Testing across AI platforms is essential: Each assistant behaves differently. Prompt testing across ChatGPT, Perplexity, Copilot, and SGE revealed that some content performed better on certain platforms. Continuous monitoring allowed the team to refine content for maximum reach.

How This Applies to Other Brands

The strategies used by Chemours can be adapted across industries:

• Software‑as‑a‑Service (SaaS) companies should identify high‑intent questions (e.g., “best project management tool for startups”) and create comparison pages with clear definitions, pricing matrices and unique differentiators. Schema markup such as SoftwareApplication, Review and FAQ will help AI assistants interpret features and user feedback.
• E‑commerce retailers can develop buyer guides and FAQs around product categories, focusing on questions that drive AI recommendations (e.g., “how to choose running shoes” or “best gifts for gamers”). Structured data like Product, Offer and AggregateRating clarifies options for AI systems.
• Local services such as clinics, salons or restaurants should maintain accurate LocalBusiness schema and answer localised queries like “best barber in [city]” or “top-rated physiotherapist near me.” Detailed service descriptions, price ranges and high‑quality photos support generative reasoning.
• Publishers and content creators need to ensure their articles start with strong definitions and context. They should use Article schema and break content into clear segments with headings and FAQs. Linking to credible sources and updating older articles will improve AI citations.

For businesses currently invisible in AI search, the first step is a content audit to identify pages lacking clear summaries and structured data. Next, map the questions users ask in AI tools and align content to answer them. Implement schema markup and ensure naming conventions, addresses, and product names are consistent across the site. Finally, monitor AI visibility regularly and refine content based on which queries drive citations.

Conclusion

Chemours’ transformation from fragmented digital presence to dominant AI answer is a testament to the power of generative engine optimisation. By unifying their web properties, strengthening entity clarity, and designing answer‑first content, they moved from minimal AI mentions to being cited in over 80% of relevant AI queries. The resulting surge in pipeline and revenue shows that AI search visibility drives real business outcomes. As AI assistants replace traditional search listings with single recommendations, brands that adapt early will capture the attention and trust of future customers. GEO and AEO are no longer optional; they are the new frontier of digital marketing. Businesses that measure and improve their AI visibility will shape tomorrow’s default choices.