D2C vs Traditional India: Who's AI-Ready? (Study)

Why we compared the two Indias

Our pilot study of 25 D2C sites found something odd: 100% had an llms.txt file and none of them wrote it. That raised the real question — if platforms are doing D2C's AI-readiness automatically, how far behind is traditional India: the clinic, the CA firm, the manufacturer whose customers live on WhatsApp and Google? So we checked, properly. We compiled 231 candidate domains, dropped everything unreachable, bot-blocked, parked or invalid (38 sites — documented, not hidden), and ran 11 live HTTP checks against each of the remaining 193 on a single day: 88 online-first D2C brands and 105 traditional businesses across healthcare, professional services, manufacturing, education and travel. The full study with methodology and raw data is open under CC-BY — every number below is verifiable down to the individual site.

The 11 checks, in plain words: does the site have an llms.txt file, and who generated it; does robots.txt block GPTBot, Google-Extended or PerplexityBot; is there any JSON-LD structured data, and specifically FAQPage and Organization/LocalBusiness blocks; is the homepage title within Google's display limit; is there a meta description, an Open Graph image, a mobile viewport tag; and is a sitemap declared in robots.txt. None of this requires special access — every check is a plain HTTP request you could run yourself from any laptop, which is exactly why we publish the per-site data instead of asking you to trust us.

Finding 1 — the 57-point llms.txt gap (that neither side earned)

90% of D2C sites serve an llms.txt vs 33% of traditional businesses. Sounds like D2C is winning the AI race — until you look at who wrote the files. 68% of all D2C sites carry a file with an explicit auto-generation signature — overwhelmingly Shopify's boilerplate, which the platform now ships by default. On the traditional side it's 14% of sites, generated by WordPress SEO plugins like All in One SEO, Rank Math and Yoast, which quietly added the feature in recent versions. India became 'AI-ready on paper' through software updates, not strategy.

The platform split explains almost everything: 76 of the 88 D2C brands run on Shopify, while traditional businesses run on custom builds (54), WordPress (26) and WooCommerce (21) — stacks where llms.txt only appears if someone installs and updates the right plugin. The llms.txt gap isn't a mindset gap. It's a platform-defaults gap wearing an AI-strategy costume.

The takeaway cuts both ways: if you're a traditional business, you're probably missing the file entirely. If you're a D2C brand, you have a generic one that says nothing distinctive about you — your bestsellers, your story, your service area are all absent. Either way, the free llms.txt generator writes a proper one in minutes.

Finding 2 — FAQ schema is where nobody is winning

Schema.org FAQPage markup — the structured data most associated with being quoted in AI answers — appears on 6% of D2C sites and 9% of traditional sites. Meanwhile 82% and 70% respectively have SOME schema, and roughly seven in ten sites in both cohorts carry an Organization or LocalBusiness block, because platforms and themes add the automatic basics. The pattern across all 193 sites is unmistakable: everything automatic is done; everything that requires sitting down and writing out your customers' actual questions is not.

And the handful of exceptions prove it's doable: a few hospital chains, dental networks and travel companies in our traditional cohort maintain what look like genuinely deliberate AI-facing files and marked-up FAQ pages. They are the outliers today — which means FAQPage schema is the cheapest differentiation left in Indian search, sitting unclaimed in both cohorts. Early movers get quoted; everyone else gets summarised.

Finding 3 — nobody is blocking AI

Across all 193 sites, blocking AI crawlers is statistical noise: 1-2% block GPTBot, zero block PerplexityBot. Contrast that with the West, where major publishers have spent two years walling out AI crawlers over licensing fights. Indian business websites are wide open — every AI engine can read essentially every Indian business site today. That means access is not the battle here. Being worth quoting when the engines arrive is.

Finding 4 — half of traditional India looks broken on WhatsApp

Only 50% of traditional business sites have an Open Graph image, vs 77% of D2C. In practice: when a clinic's or CA firm's link is shared on WhatsApp — India's primary business channel — half the time it shows a bare grey preview with no image, which reads as 'broken' or 'suspicious' to the person receiving it. The supporting basics lean the same way: 87% of traditional sites have a meta description (vs 97% D2C) and only 73% declare a sitemap in robots.txt (vs 95%). This is the most fixable finding in the study — the meta tags generator solves the preview problem in under an hour.

Finding 5 — traditional businesses beat D2C on titles

The one result that surprised us: 50% of traditional sites keep their homepage title within Google's ~60-character display limit, vs only 41% of D2C brands. D2C titles overflow with offer text — 'Buy X Online | Best Y in India | Free Shipping | 20% Off' — while the average CA firm or manufacturer states its name and what it does. Restraint, it turns out, is a ranking asset: an overflowing title gets truncated into something Google chooses for you.

The study that corrected itself

One more thing, because honest research should show its work: our same-day 25-site pilot said 100% of D2C sites have llms.txt and 12% have FAQ schema. At 88 sites those numbers fell to 90% and 6%. Same method, same day — the pilot simply sampled the most-listed, most search-visible brands, and visibility correlates with platform polish. We publish the correction rather than bury it, and it carries a warning for every 'top brands' study you read (including ours): listicle sampling flatters the numbers. Real-world percentages across all of India are almost certainly lower than anything reported here.

What to do about it, by Monday

• Traditional business: add an llms.txt (the free generator does it), and fix your link preview with the meta tags generator — both are sub-hour jobs.
• D2C brand: replace the boilerplate llms.txt with one that describes YOUR brand, and trim your title tags — 59% of your cohort overflows Google's display limit.
• Everyone: write real FAQPage schema from questions customers actually ask — the Schema Generator handles the markup. At 6-9% adoption, early movers get quoted.
• Check where you stand first: the free scan runs these exact checks on your site in about a minute, no signup.

The honest caveats

Cohorts were sampled from public 'top brands/firms' listicles plus our own tracking — that skews toward businesses already visible in search, so all-India percentages are likely LOWER, especially for traditional businesses. 38 unreachable or bot-blocked candidates were excluded and are listed as exclusions, not silently dropped. And llms.txt authorship can't be proven from outside — we classify only by explicit generation signatures and report everything else as unattributed, not 'hand-written'. Full limitations, the per-site table and the CC-BY raw dataset are on the study page.

Frequently asked questions