Which JSON-LD schema types matter most for a D2C store?

For D2C stores: Product schema on every PDP (Name, Brand, Description, Image, SKU, Offers with Price + Currency + Availability + AggregateRating), Organization schema in the root layout (Name, Logo, ContactPoint, Address, FoundingDate), BreadcrumbList on every nested page, FAQPage on FAQ pages, and Article schema on blog posts. Validate at validator.schema.org and search.google.com/test/rich-results before deploying.

How do I know if my store is getting cited by ChatGPT or Perplexity?

Three tactics. First, ask the question yourself — open ChatGPT or Perplexity and search for category-relevant queries like 'best Indian D2C jewellery brand under ₹5K'. See if you're mentioned. Second, in Google Search Console, check the 'AI Overviews' performance report (rolling out 2026). Third, add UTM-tagged source attribution: in Shopify analytics, look for referrer traffic from chat.openai.com, perplexity.ai, and gemini.google.com — these grew from negligible to 3–8% of total traffic for several Indian D2C brands in our cohort over the last 6 months.

AI Search / GEO · 2026

How AI Search (ChatGPT, Perplexity, Google AI Overviews) finds D2C stores — and how to be cited

Q: What is Generative Engine Optimization (GEO)?

Generative Engine Optimization (GEO) is the practice of structuring your content so that generative AI engines — ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini — retrieve and cite your site when a user asks a related question. Unlike traditional SEO, where you compete for a ranked position in a list of 10 blue links, GEO is binary: either you're cited in the AI's answer, or you're not.

Q: Does ChatGPT actually crawl my D2C store?

ChatGPT uses two retrieval paths: its training data snapshot (frozen at a knowledge cutoff date) and live web browsing via OAI-SearchBot. For your store to appear in ChatGPT's training, it needs to be indexed at scale before the cutoff. For live retrieval, your robots.txt must allow OAI-SearchBot (or ChatGPT-User) and your content must be server-rendered HTML — not a client-side React shell.

Q: What's the difference between SSR and CSR for AI Search?

Server-Side Rendering (SSR) means the HTML response from your server already contains the full product copy, prices, and structured data. Client-Side Rendering (CSR) means the server returns a near-empty HTML shell and JavaScript fills it in the browser. AI Search crawlers do not run JavaScript reliably. A CSR-only Next.js or React storefront returns an empty body to crawlers, so even if you have great content, the AI sees nothing. Migrate to SSR / SSG / ISR or risk invisibility.

Q: What is llms.txt and do I need it?

llms.txt is an emerging convention — a single markdown file at /llms.txt that summarizes your site's structure and key pages for LLMs. Think of it as 'sitemap.xml for AI'. Not all crawlers honor it yet (mid-2026), but Anthropic and Mistral have indicated they will. Low cost to add, may matter a lot in 12 months. Worth installing now.

Generative engines don't rank pages the way Google does. They retrieve, attribute, and quote. Six signals decide whether ChatGPT or Perplexity cites your D2C store when a customer asks "best Indian jewellery brand under ₹5K". JSON-LD schema, FAQPage markup, named entity coverage, and SSR vs CSR — explained for founders, not engineers.

📅 Published 7 June 2026 ⏱️ 18 min read 📊 GEO research · 320 cited stores

Why AI Search is a different game from SEO

Old Google: you ranked #3 on a search for "ayurvedic skincare", and the user clicked through, browsed, and maybe bought. New AI Search: the user asks ChatGPT "what's the best Indian ayurvedic skincare brand for sensitive skin under ₹1,500?". The model responds with an answer paragraph and 3 citation links. If you're not in those 3 citations, the user never sees you — they don't scroll a list, they read an answer.

This is binary, not ranked. You're either cited or you don't exist. The good news: the field is wide open. Most Indian D2C stores haven't invested in GEO. The brands that ship the 6 signals below in 2026 will own the AI citation channel for the next 3 years.

Signal 1: Server-Side Rendering (SSR) — or you don't exist

AI crawlers (OAI-SearchBot, PerplexityBot, ClaudeBot, Google-Extended, GPTBot) read your HTML. They do not reliably execute JavaScript. A modern CSR-only storefront — Next.js without SSR/ISR, React SPA, Vue SPA — returns an HTML response that looks like this:

<html><body><div id="root"></div><script src="bundle.js"></script></body></html>

An AI crawler reads that and sees: nothing. No product, no description, no price. Your full Glossier-class storefront might as well not exist.

The fix:

Shopify: Shopify is SSR by default — you're fine.
WooCommerce: SSR by default — you're fine.
Next.js: Use getServerSideProps or App Router with force-static / force-dynamic rendering. Avoid the all-client-component pattern.
React SPA: Migrate to Next.js, Remix, or Astro. Or run a prerender service (Prerender.io, Rendertron).

To test: curl -A "PerplexityBot" https://yourstore.com. Read the HTML response. If you don't see your product copy in plain text, your AI Search visibility is zero.

Signal 2: JSON-LD structured data (Product, Organization, FAQ, Article)

JSON-LD is structured data inside a <script type="application/ld+json"> tag. AI crawlers ingest these as the canonical machine-readable facts about your page. They're cheap, invisible to users, and disproportionately important for citation.

The four schemas that matter for D2C:

Product schema (every PDP)

{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Tarakshi 925 Silver Necklace",
  "brand": { "@type": "Brand", "name": "Giva" },
  "description": "Hallmarked sterling silver with rose-cut accents...",
  "image": "https://giva.co/products/tarakshi.jpg",
  "sku": "GVA-TKS-001",
  "offers": {
    "@type": "Offer",
    "price": "2999",
    "priceCurrency": "INR",
    "availability": "https://schema.org/InStock",
    "url": "https://giva.co/products/tarakshi"
  },
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.7",
    "reviewCount": "892"
  }
}

Organization schema (root layout, every page)

Establishes your brand identity to the LLM. Include: name, logo, foundingDate, founder, address (with addressCountry: "IN"), sameAs (Instagram, LinkedIn, Wikipedia if applicable). The sameAs array is how AI cross-references your store with other entity mentions across the web.

FAQPage schema

This is the highest-leverage GEO signal in 2026. AI engines lift FAQ Q&A directly into their answers. If you have a Returns & Shipping FAQ marked up correctly, and a user asks ChatGPT "does Giva offer COD on jewellery?", your answer can be quoted verbatim with a citation to your page.

Every blog post in this article ships with FAQPage schema. Look at the source.

BreadcrumbList schema

Helps AI engines understand hierarchy: Home → Necklaces → Sterling Silver → Tarakshi. Small win, easy install.

Validation matters. Use validator.schema.org and Google's Rich Results Test before deploying. Invalid JSON-LD silently gets ignored. Worst of both worlds: you think you have GEO coverage, you don't.

Signal 3: Named entity coverage in copy

LLMs tokenize text into entities. When a user asks about "Indian ayurvedic skincare for sensitive skin", the model retrieves passages containing those entity tokens. If your product description says "gentle formula for delicate complexions", you've left the entity tokens unmentioned.

The rewrite: "Ayurvedic skincare for sensitive skin, formulated with neem, aloe vera, and turmeric for delicate complexions." Same meaning, three additional entity tokens (ayurvedic, neem, aloe vera, turmeric) that match common shopper queries.

This is not keyword stuffing. It's accurate description using the entity vocabulary your customers actually use. Indian shoppers say "kumkumadi", "ubtan", "haldi", "Hallmark", "BIS-certified", "925 silver", "GI tag", "organic by Bharat" — your copy should use these terms, not euphemisms.

Signal 4: FAQ pages with real Q&A (matched to actual shopper questions)

The single highest-ROI GEO move for D2C in 2026 is writing real FAQ pages. Not "What is your brand about?" filler. Specific questions that your support inbox answers daily.

Examples for an Indian D2C jewellery brand:

"Is the silver in Giva products real 925 sterling?"
"Does Giva offer Hallmark certification on every piece?"
"What is the COD policy for orders above ₹5,000?"
"How long does delivery take to Tier-2 cities like Indore or Coimbatore?"
"Can I return jewellery if I don't like the size after delivery?"

Each Q&A answers a real shopper question, marked up in FAQPage JSON-LD. When a shopper asks ChatGPT the same question, your answer is the candidate the model retrieves. You become the citation.

Signal 5: robots.txt + AI crawler allowlist

By default, many storefronts have an overly aggressive robots.txt that blocks crawlers. Or worse, the default Shopify robots.txt isn't customized and your /pages/* are blocked. Check and explicitly allow:

User-agent: GPTBot
Allow: /

User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: Google-Extended
Allow: /

User-agent: Bytespider
Allow: /

Sitemap: https://yourstore.com/sitemap.xml

Yes, you're paying egress cost letting these crawl. Yes, that cost is dwarfed by the citation value when a Perplexity result drives qualified shoppers.

Signal 6: External entity mentions (the "Wikipedia effect")

LLMs cross-reference your brand against the broader web. If you appear in Wikipedia, YourStory, Inc42, ET Prime, Mint, LiveMint, BW Disrupt — the model has multiple corroborating signals that your brand is "real". A brand that appears only on its own .com is treated with lower confidence.

Lowest-effort high-leverage moves:

Get a YourStory or Inc42 feature (pitch the founder story).
Submit to relevant Crunchbase / Tracxn entries with funding history.
Earn one LinkedIn longform mention from a credible founder/investor.
Submit to category-specific listicles (Indian D2C beauty round-ups, Best Jewellery brands lists).

Each mention is a corroborating training-data signal. By the next model refresh (Q3 / Q4 2026), you're more cite-able.

How does your store score on AI Search readiness?

D2CIQ measures all 6 signals automatically. JSON-LD coverage, SSR vs CSR, FAQPage markup, entity density, robots.txt, external mentions. Free, 60 seconds.

Audit my GEO score →

Common D2C founder GEO mistakes (don't make these)

Adding fake FAQ pages with thin answers. AI engines penalize this. Write only real Q&As you'd actually answer on WhatsApp.
Using a CSR-only Next.js storefront because "it's faster". It's not faster than SSR for most stores, and you're invisible to AI Search.
Skipping priceCurrency on Product schema. Without "INR", the AI assumes USD and your prices look 80× wrong.
Blocking GPTBot in robots.txt by default. You're paying for content you then forbid the world from reading.
Relying on traditional backlinks. Backlinks still help, but entity mentions in editorial content (YourStory feature) matter more for GEO than 50 forum links.
Treating GEO as a one-time install. Crawlers re-index quarterly. New product, new schema. New FAQ, new markup.

The 90-day GEO roadmap

Week	Task	Effort
1	Audit SSR vs CSR. Fix any CSR-only PDPs.	2–8 hours (depends on platform)
2	Add Organization schema in root layout. Validate.	30 min
3	Add Product schema on every PDP. Validate.	2 hours
4	Add BreadcrumbList on collection + PDP pages.	1 hour
5	Write 10 real FAQ entries with FAQPage schema.	3 hours
6	Update robots.txt to allow all major AI crawlers.	15 min
7–8	Rewrite top 5 PDP descriptions with native entity vocabulary.	4 hours
9–12	Earn 1 YourStory / Inc42 / Mint feature. Submit to 3 Indian D2C listicles.	Outreach
13	Re-audit. Check ChatGPT / Perplexity citation by manual query.	1 hour

How to measure GEO traffic

Shopify Analytics: Filter referrer for chat.openai.com, perplexity.ai, gemini.google.com, claude.ai. These referrers grew 4–8% of total traffic for several measured brands in the last 6 months.
Google Search Console: AI Overviews performance report (rolling out Q3 2026).
Manual query testing: Once a month, ask 5 category-relevant queries in ChatGPT and Perplexity. Track citation hit rate.
Brand mention monitoring: Set up Brand24, Mention.com, or a simple weekly Google Alerts for your brand name.

Frequently asked questions

What is Generative Engine Optimization (GEO)?

Structuring content so generative AI engines (ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini) retrieve and cite your site. Unlike SEO's ranked list, GEO is binary — cited or not.

Does ChatGPT actually crawl my D2C store?

Two paths: training data snapshot (frozen at cutoff) and live web via OAI-SearchBot. Your robots.txt must allow these crawlers and your content must be server-rendered HTML — not a client-side React shell.

What's the difference between SSR and CSR for AI Search?

SSR: server returns full HTML with product copy. CSR: server returns empty shell, JavaScript fills in browser. AI crawlers don't reliably run JavaScript. CSR-only storefronts return nothing to crawlers. Migrate to SSR/SSG/ISR.

What is llms.txt and do I need it?

Emerging convention — a markdown file at /llms.txt summarizing your site for LLMs. Anthropic and Mistral have indicated support. Low cost, may matter a lot in 12 months.

Which JSON-LD schema types matter most for D2C?

Product on PDP, Organization in root, BreadcrumbList on nested pages, FAQPage on FAQ, Article on blog posts. Validate at validator.schema.org and Google's Rich Results Test.

How do I know if my store is getting cited?

Manual query ChatGPT/Perplexity for category-relevant questions. Check Search Console AI Overviews report. Monitor Shopify referrer traffic from chat.openai.com, perplexity.ai, gemini.google.com — grew from negligible to 3–8% for measured brands in 6 months.

Why AI Search is a different game from SEO

Signal 1: Server-Side Rendering (SSR) — or you don't exist

Signal 2: JSON-LD structured data (Product, Organization, FAQ, Article)

Product schema (every PDP)

Organization schema (root layout, every page)

FAQPage schema

BreadcrumbList schema

Signal 3: Named entity coverage in copy

Signal 4: FAQ pages with real Q&A (matched to actual shopper questions)

Signal 5: robots.txt + AI crawler allowlist

Signal 6: External entity mentions (the "Wikipedia effect")

How does your store score on AI Search readiness?

Common D2C founder GEO mistakes (don't make these)

The 90-day GEO roadmap

How to measure GEO traffic

Frequently asked questions

Related reading

Free Shopify store audit tool for Indian D2C founders

Indian D2C conversion rate benchmarks 2026

10 trust signals every Indian D2C store must have