How a Singapore DTC Brand Increased AI Search Citations by 3.2x in 90 Days

From Invisible to Top 3 in AI Recommendations

When a DTC brand we work with first came to us, they had a solid Shopify store, decent Google traffic, and a loyal customer base. But when we tested their AI visibility by asking ChatGPT, Claude, and Perplexity to recommend products in their category, the brand did not appear anywhere. Not in the top five. Not in the top ten. Not at all.

Ninety days later, the same queries returned the brand as a top three recommendation across all three major AI models. AI referral traffic became a meaningful acquisition channel, and the brand's overall citation count increased by 3.2 times.

This is the story of how we did it, using the same five-dimension framework we apply to every Swop Labs AI Search engagement. If you are unfamiliar with AEO fundamentals, our guide to AI search optimisation provides the foundation.

The Starting Point: A D Grade Audit

We began with our standard five-dimension AI visibility audit. The results were sobering.

Structured data scored poorly. The Shopify store had basic product markup but was missing Organization, LocalBusiness, FAQPage, and BreadcrumbList schemas. The existing product markup was incomplete, lacking review aggregation and detailed attribute data.

Content authority was minimal. The store had product pages and a brief about page, but no blog, no guides, and no educational content. There was nothing to signal topical expertise to AI models.

Review signals were moderate. The brand had Google reviews but they were not aggregated into structured data. Reviews on the Shopify store itself were not exposed in a way that AI models could access.

The citation network was weak. Beyond the brand's own website and social media profiles, there were very few third-party mentions on authoritative sites. No media coverage, no directory listings beyond the basics, and no industry publications.

AI-specific metadata was non-existent. No llms.txt file, no agent.json, and no optimisation for conversational queries.

The overall grade: D. The brand was effectively invisible to AI models.

The 90 Day Activation

Weeks 1 to 2: Foundation

The first priority was fixing the structured data. We implemented comprehensive Schema.org markup across the entire site. This included Organization and LocalBusiness schemas with complete business details, enhanced Product markup with detailed attributes, pricing, availability, and review aggregation, FAQPage schema on every page that contained Q&A content, BreadcrumbList for site navigation, and a new llms.txt file at the root domain.

We also created an agent.json file, a machine-readable product catalogue designed specifically for AI model consumption. This gave AI models a structured way to understand the brand's full product range, positioning, and unique selling points.

Weeks 3 to 6: Content Authority

The brand had zero educational content. We built a content cluster of twelve articles covering the brand's core product category: buyer guides, comparison articles, how-to guides, care instructions, and industry explainers.

Each article was written to answer specific questions that consumers ask AI models. Instead of keyword-optimised SEO copy, we wrote in natural, conversational language that matched the way people phrase queries to ChatGPT and Claude.

The articles linked to each other extensively, creating a topical web that signalled deep expertise. They also linked to the relevant product pages, establishing a clear relationship between educational content and commercial offerings.

Weeks 4 to 8: Citation Building

We identified authoritative sites where the brand should be mentioned but was not. This included industry directories, Singapore business listings, review platforms, and relevant publications.

We secured mentions through a combination of outreach, guest contributions, and strategic partnerships. Each mention included consistent business information (name, website, description) to strengthen the citation network.

We also optimised the brand's Google Business Profile, which many ecommerce brands neglect. Even though the business is primarily online, having a complete GBP with reviews, products, and regular updates contributes to AI citation signals.

Weeks 6 to 10: Review Amplification

The brand had genuine customer reviews, but they were scattered and not well structured. We implemented a systematic review collection process that increased the volume of Google reviews significantly. We also ensured all reviews were aggregated into the site's structured data so AI models could access them.

Weeks 8 to 12: Monitoring and Optimisation

Starting in week eight, we began running daily citation checks across ChatGPT, Claude, and Perplexity for the brand's target queries. This allowed us to see which optimisations were having the most impact and adjust our approach in real time.

We noticed that Claude responded fastest to structured data improvements, while ChatGPT placed more weight on third-party mentions and review signals. Perplexity's citations correlated most closely with content authority. These insights shaped our prioritisation for the final weeks.

The Results

By day 90, the brand's AI visibility had transformed.

The five-dimension audit score moved from D to A. Structured data: A. Content authority: B+. Review signals: A. Citation network: B. AI-specific metadata: A.

The brand appeared in the top three recommendations for seven out of ten target queries across ChatGPT, Claude, and Perplexity. Before the engagement, it appeared in zero.

Total AI citations increased by 3.2 times, measured across weekly spot checks of 20 target queries per model.

AI referral traffic, tracked through UTM parameters and referral analysis, grew from negligible to a meaningful percentage of overall acquisition. The cost per acquisition from AI referrals was lower than paid advertising, making it one of the brand's most efficient channels.

What This Means for Your Business

The specific tactics we used for this brand are confidential, as they are part of our proprietary methodology. But the framework is universal. Any Singapore business can improve its AI visibility by addressing the same five dimensions: structured data, content authority, review signals, citation network, and AI-specific metadata.

The key insight is that AI visibility is not random. It is a direct result of specific, measurable signals that you can influence. Businesses that invest in these signals now will capture the AI recommendation positions. Businesses that wait will find it increasingly difficult to break in as their competitors strengthen their positions.

For a step-by-step guide to improving your own ChatGPT visibility, read our ChatGPT citation guide. For a broader understanding of AEO, start with our complete guide to AI search optimisation.

Get Your AI Visibility Audit

The Swop Labs AI Search team runs the same five-dimension audit for every prospective client. It shows you exactly where you stand and what to prioritise.

Chat with us on WhatsApp to request yours. We reply within one Singapore business day.