India's largest digital-first mattress brand isn't AI-ready.
We ran the GEO Score scanner on 6 Indian sleep and mattress brands — the same 15 signals, the same tool anyone can use for free. Wakefit — ₹1,274 Cr revenue in FY25, India's first listed sleep-tech company — scored 38 out of 100. The Sleep Company — ₹500 Cr revenue, 160 stores — scored 87. That's a 49-point gap between the category's largest digital-first brand and a competitor less than half its size.
The brand everyone knows is the brand AI can't discover.
The full leaderboard
| Brand | GEO Score | Band | Accessibility | Readability | Understandability |
|---|---|---|---|---|---|
| The Sleep Company | 87 | Fully Optimised | 33/33 | 28/37 | 26/30 |
| Duroflex | 84 | Fully Optimised | 32/33 | 29/37 | 23/30 |
| Kurl-On | 79 | Well Optimised | 24/33 | 28/37 | 27/30 |
| SleepyCat | 78 | Well Optimised | 33/33 | 23/37 | 22/30 |
| Flo Mattress | 68 | Well Optimised | 25/33 | 25/37 | 18/30 |
| Wakefit | 38 | Not Optimised | 11/33 | 21/37 | 6/30 |
Category average: 72/100. Five of six brands score 68 or above. Then there's Wakefit at 38 — the only "Not Optimised" brand in the category.
The Sleep Company and Duroflex lead as the two "Fully Optimised" brands. Kurl-On and SleepyCat cluster near 79 — well optimised, with different strengths. SleepyCat has perfect Accessibility (33/33) but weaker Readability. Kurl-On is the reverse — strongest Understandability (27/30) in the category but weaker Accessibility because there's no llms.txt. Flo Mattress sits at 68 with solid fundamentals.
And Wakefit? Here's the paradox: the content is actually decent. Readability scores 21/37 — the homepage has 1,618 words, 8 headings, and 7 JSON-LD blocks. The problem isn't what's on the site. It's that AI crawlers can't find any of it.
Why Wakefit scores 38
This isn't a content problem. It's a discoverability problem.
Wakefit's robots.txt returns HTTP 403. Not a permissive file, not a restrictive file — a flat rejection. The server won't even tell crawlers what's allowed. Their llms.txt? Also 403. Their sitemap? Also 403. No about page detected. No author attribution on any content.
The content itself is fine. When a regular browser visits Wakefit's homepage, it finds 1,618 words, 8 headings, and 7 JSON-LD structured data blocks. The Readability score is a respectable 21/37. But the Accessibility score — the pillar that measures whether AI crawlers can discover your content — is just 11/33. The front door is locked. The content behind it doesn't matter if nobody can get in.
Here's the Accessibility breakdown across all six brands:
| Brand | Accessibility | robots.txt | Sitemap | Crawlability | llms.txt |
|---|---|---|---|---|---|
| SleepyCat | 33/33 | 200 OK, GPTBot explicitly allowed | Full sitemap | 5 pages crawled | 366,778 chars |
| Duroflex | 32/33 | 200 OK, GPTBot explicitly allowed | Full sitemap | 5 pages crawled | 3,764 chars |
| Flo Mattress | 25/33 | 200 OK, all crawlers allowed | Full sitemap | 5 pages crawled | 404 — not found |
| The Sleep Company | 33/33 | 200 OK, all crawlers allowed | Partial (503 on some sub-sitemaps) | 5 pages crawled | 119,702 chars |
| Kurl-On | 24/33 | 200 OK, all crawlers allowed | Full sitemap | 5 pages crawled | 404 — not found |
| Wakefit | 11/33 | HTTP 403 | HTTP 403 | 1 page crawled | HTTP 403 |
This almost certainly isn't intentional. A Cloudflare WAF rule — likely a blanket block on automated requests — is intercepting crawler traffic before it reaches the origin server. Wakefit's engineering team probably doesn't know this is happening. The WAF is doing its job protecting against bots. The problem is that AI crawlers are now the bots you want to let through.
The result: when someone asks ChatGPT or Perplexity "best mattress brand in India", Wakefit's website contributes nothing to the answer. Whatever AI knows about Wakefit comes from third-party sources — reviews, news articles, competitor comparisons. Wakefit has zero control over how it's represented.
Why The Sleep Company leads at 87
Want to know how your brand scores on these same metrics?
We'll run 20 prompts across 3 AI platforms and send your report within 24 hours.
The Sleep Company is doing everything right across all three pillars. Perfect Accessibility (33/33), strong Readability (28/37), and solid Understandability (26/30).
Accessibility: 33/33. llms.txt is 119,702 characters — a comprehensive, structured document that tells AI models exactly what the brand is about, what products they sell, and what makes their SmartGRID technology different. Robots.txt allows all crawlers. Full sitemap accessible. Five pages crawled without friction.
Readability: 28/37. Correct heading hierarchy across most pages. FAQ sections with structured markup on product pages — 10 FAQ pairs on the mattress page alone. When someone asks "is The Sleep Company mattress good for back pain?", the FAQ schema gives AI models a pre-formatted question-answer pair to pull from. Blog content averages 2,400+ words with named authors.
Understandability: 26/30. Organization, FAQPage, and BreadcrumbList JSON-LD schema on key pages. Named authors with proper markup — Shubh Madhyan, Nishant Gosavi — across blog content. AI models treat authored content as more credible than anonymous brand copy.
The Sleep Company isn't just passively crawlable — they've actively structured their content for AI consumption. That's the difference between leaving the door open and putting up a welcome sign with directions.
The SleepyCat story: perfect Accessibility, weak Readability
SleepyCat is the most interesting case in this scan. They score 78 — solidly Well Optimised. Their Accessibility is a perfect 33/33. They have an llms.txt file that's 366,778 characters — one of the largest we've ever seen. Robots.txt explicitly allows GPTBot. Someone at SleepyCat clearly understands that AI crawlers need access and structured content.
But their Readability is 23/37 — the weakest pillar.
They opened the door wide but the content behind it isn't structured for AI to parse.
Heading hierarchy has jumps (H2 to H4), the blog index is missing an H1, and meta descriptions are either too short (32 characters on the homepage) or missing. The content is there — SleepyCat has good product pages and a substantial blog. The issue is structural: headings, answer blocks, and metadata aren't optimised for how AI models extract and understand information.
The fix: restructure heading hierarchy, add FAQ markup to product pages, and fix meta descriptions. Their Accessibility foundation is already perfect. Fixing Readability could push them from 78 into the mid-80s.
The bigger pattern: revenue doesn't predict AI-readiness
Here's what makes this scan significant beyond mattresses.
We've now scanned 48 brands across 8 categories — travel and luggage, skincare, audio, CRM, HR, accounting, men's grooming, and now sleep. The pattern is consistent: the highest-revenue brand in a category is rarely the most AI-ready.
Revenue and AI-readiness are decoupled. GEO Score measures whether your website is structured for AI to access, parse, and understand — not whether AI actually recommends you. But if AI can't read your site, it can't use your content as a source when answering questions.
This maps to the three GEO Score pillars directly. Accessibility — can AI crawlers reach your content? Readability — is your content structured for AI to parse and extract? Understandability — can AI determine meaning, credibility, and authority from your content?
Wakefit likely has strong brand authority from years of press coverage, funding news, and customer reviews across the web. Their content is actually decent (Readability: 21/37). But their Accessibility is broken (11/33) — robots.txt, sitemap, and llms.txt all return 403. AI crawlers can't discover the content that's there. All that brand equity goes uncompounded because the discovery layer is locked.
The Sleep Company has all three pillars working. Crawlers get in (Accessibility: 33/33). Content is structured with FAQs, headings, and llms.txt (Readability: 28/37). And the brand has authored content with proper markup (Understandability: 26/30).
The 49-point gap between them isn't about who makes a better mattress. It's about who built their website for the way AI discovers and processes information.
What Wakefit could fix this week
Three changes. Days, not months. No content strategy overhaul required.
1. Whitelist AI crawler user agents in Cloudflare WAF.
Go to the Cloudflare dashboard. Find the WAF rule that's returning 403 to automated requests. Add exceptions for GPTBot, ClaudeBot, PerplexityBot, Google-Extended, and Bytespider. This single change would fix the robots.txt 403, make the sitemap accessible, and let crawlers actually read the site's content — pushing Accessibility from 11/33 toward 25+.
2. Add an llms.txt file.
A single text file at wakefit.co/llms.txt that describes the brand, product lines, materials, certifications, and key differentiators. The Sleep Company's is 119K characters. Wakefit could start with 5,000 characters covering the basics. Our llms.txt guide for Indian brands has the exact format.
3. Add an about page with author attribution.
Wakefit has no detectable about page and no author attribution on any content. Add a proper /about page with founder bios, company history, and team credentials. Add author bylines with schema markup to blog posts and buying guides. AI models weigh authored content from identified experts higher than anonymous brand copy.
Conservative estimate: these three fixes move Wakefit from 38 to 65+. That's a jump from "Not Optimised" to "Well Optimised" without writing a single new blog post or changing any product copy. Pure infrastructure.
Run the free GEO Score scanner on your own site to see where your Accessibility, Readability, and Understandability pillars stand.
The 49-point gap isn't destiny. It's a configuration file.
Frequently asked questions
Which Indian mattress brand is most AI-ready?
The Sleep Company leads with a GEO Score of 87/100 (Fully Optimised), followed by Duroflex (84), Kurl-On (79), and SleepyCat (78). Wakefit scores lowest at 38/100 — Not Optimised — dragged down by Accessibility issues, not content quality.
Why does Wakefit score so low on AI-readiness?
Wakefit's Cloudflare WAF returns HTTP 403 to robots.txt, sitemap, and llms.txt requests. The content itself is decent (Readability: 21/37), but AI crawlers can't discover it because the infrastructure is blocked. Accessibility scores just 11/33. This is a WAF configuration issue — not an intentional decision.
What is a GEO Score?
A GEO Score is a 0–100 AI-readiness metric that measures how well a website is structured for AI crawlers and citation engines. It checks 15 signals across 3 pillars: Accessibility (can AI crawlers reach your content?), Readability (is content structured for AI to parse?), and Understandability (can AI extract meaning and authority?).
Can Wakefit fix its AI-readiness quickly?
Yes. Three infrastructure fixes could move Wakefit from 38 to 65+: whitelist AI crawler user agents in Cloudflare WAF settings to fix the robots.txt/sitemap/llms.txt 403 errors, add an llms.txt file describing the brand, and add an about page with author attribution. These are Accessibility fixes that take days, not months.