AEO / GEO Audit
Provogue (provogue.com) — AI SEO / AEO Audit
Report generated: 2026-08-11. Domain: https://www.provogue.com/. Method: live technical checks (robots.txt, llms.txt, agents.md, UCP, raw-HTML schema fetch, policy-link status, H1/heading checks, blog recency) + web-search ranking-presence proxy for priority queries. Business context from business-analysis.md.
0. TL;DR (what matters)
Provogue is ahead of 95% of Indian D2C stores on the agentic front — it ships llms.txt, agents.md, a full Universal Commerce Protocol (UCP) merchant profile, and an agentic-discovery sitemap. That is rare and positions it well for the emerging "AI agents buying on your behalf" wave.
But on the classic AI-visibility axis (getting cited by ChatGPT / Perplexity / Gemini / Google AI Overviews for the queries that drive high-intent traffic), Provogue is weak:
- It wins only branded queries ("provogue luggage"). It is absent from every "best luggage brand / luggage comparison" answer we tested — exactly the high-intent, high-consideration queries where competitors (Mokobara, Safari, VIP, Nasher Miles, Uppercase, NORI) get cited.
- On-page extractability has real gaps: no H1 on collection and about pages, no FAQPage schema despite a 59-question FAQ, thin PDP descriptions (≈219 words), a shipping-policy page that does not exist, and an
AggregateRatingwith only 1 review that undercuts the trust signal it should provide. pricing.md(a low-effort, high-value file for AI agents) is missing (404).
The opportunity is not "catch up on agentic infra" — it is "make the existing, rich, well-structured catalogue citable for the queries shoppers actually ask AI." The roadmap below is sequenced by impact.
1. AI Bot Access (robots.txt) — PASS
AI crawlers are NOT blocked. The storefront robots.txt has no Disallow rules for GPTBot, ChatGPT-User, PerplexityBot, ClaudeBot, anthropic-ai, Google-Extended, or Bingbot. Only transactional/private paths are disallowed (/admin, /cart/, /checkout, /orders, /account).
Implication: every major AI engine is technically able to crawl and cite Provogue. The visibility problem is therefore content/structure, not access.
2. Machine-Readable Files for AI Agents
| File | Status | Notes |
|---|---|---|
/llms.txt | 200 ✅ | Detailed; describes UCP + Shop skill; links to policies. |
/agents.md | 200 ✅ | Canonical agent-facing doc (mirrors llms.txt). |
/.well-known/ucp | 200 ✅ | Full UCP merchant profile (catalog, cart, checkout, discount, payment handlers incl. Google Pay + Shopify card). |
/api/ucp/mcp | referenced ✅ | MCP endpoint declared in UCP for agent-driven commerce. |
/sitemap_agentic_discovery.xml | 200 ✅ | Shopify auto-generated agentic sitemap present. |
/pricing.md | 404 ❌ | Missing. A structured pricing file for AI agents is absent. |
/pricing.txt | 404 ❌ | Missing. |
Findings:
- Provogue is genuinely advanced here. Most D2C brands have none of these; Provogue has all four core agentic primitives.
- Gap:
/pricing.mdis missing. While the UCP catalog endpoint exposes pricing, a plain/pricing.mdis a trivial, durable, parseable artifact that non-UCP agents (and humans pasting into ChatGPT) can read directly. The agent docs never mention/pricing.md, so this is a clean addition, not a fix.
3. Schema Markup (verified via raw HTML, not text-extraction)
Verified by fetching raw page source and parsing <script type="application/ld+json"> blocks (the method that preserves script tags — text-extraction checks would falsely report "no schema").
| Page type | Schema found | Assessment |
|---|---|---|
| Homepage | Organization, WebSite (+SearchAction) | OK. No Article/blogposting; fine for home. |
| Product (PDP) | Product + Brand + Offer + AggregateRating | Strong for non-Google engines — full Product/Offer/AggregateRating. BUT thin (see §5). |
| Collection | ItemList (+ListItem per product), Organization | Good product-list structure. No BreadcrumbList. |
| About | Organization only | OK. |
| FAQ | Organization only | Gap: 59 questions but no FAQPage schema (see §4). |
| Blog post | 5 ld+json blocks (Organization + Article-class) | Present; confirm BlogPosting/Article with author+date. |
Verified specifics (Lattitude Core Large PDP):
Offer:sku: PRPPCORESBLU28,price: 2399.0 INR,availability: InStock✅Brand: Provogue ✅aggregateRating:ratingValue 5, ratingCount 1⚠️ — only 1 review. A 5.0 from 1 review reads as untrustworthy to AI and humans alike.skupresent at Offer level; nogtin/mpnon the Product node.
4. Content Extractability — GAPS
4.1 H1 / heading structure (verified in raw HTML)
| Page | H1 count | Problem |
|---|---|---|
| Homepage | 1, but it is the logo link (<h1 class="header__logo">) | No descriptive topical H1. AI uses H1 as the primary topic signal. |
Collection (/collections/luggage) | 0 | Collection title is not wrapped in an H1. Major — collection pages are the entry point for "best luggage" queries. |
| About | 0 | No H1. |
| FAQ | H2/H3 present (59 questions) | Headings exist; schema missing (below). |
AI engines rely on H1 for topic identification. Collection and about pages having no H1 is a concrete, fixable defect that weakens topical signals for category queries.
4.2 FAQPage schema missing
The FAQ page carries 59 real Q&A pairs (returns, warranty, personalisation, shipping questions) — exactly the kind of extractable content Perplexity/ChatGPT love — but it is tagged only as Organization. Adding FAQPage JSON-LD (or at minimum well-formed <h2>+<p> pairs, already present) makes these directly extractable. Google has de-emphasised FAQ rich results for non-authoritative sites, but ChatGPT/Claude/Perplexity still reward it — and these are the engines where Provogue is weakest.
4.3 PDP description depth
The sampled PDP body copy is ≈219 words — below the 300–800-word band AI engines favour for comprehensive product extraction. With 233 SKUs on a shared template, many product descriptions are likely thin marketing copy. This is the single biggest on-page content gap for "best/what is" product queries.
4.4 AggregateRating credibility
ratingCount: 1 on a flagship product. The business analysis already flags shallow review/social proof on PDPs as a high-impact gap. Until review volume grows (Judge.me / Yotpo recommended), the AggregateRating schema is doing more harm than good (signals "unreviewed / untrusted" to extraction).
4.5 Missing BreadcrumbList schema
Collection pages emit ItemList but no BreadcrumbList. AI uses breadcrumbs to understand category hierarchy ("luggage → hard luggage → cabin"). Add BreadcrumbList to collection + PDP templates.
5. Policy Pages & Trust Signals
Footer policy links all return 200 (no broken footer links): /pages/privacy-policy, /pages/returns-cancellations, /pages/terms-conditions-1, /pages/contact, /pages/about-us. Good — broken policy links (a common AI-trust eroder) are NOT an issue here.
But two real gaps:
- No shipping / delivery policy page exists at all — tested
/pages/shipping-policy,/policies/shipping-policy,/pages/shipping,/pages/delivery→ all 404. For a pan-India D2C store, the absence of a shipping policy is both an AI trust-signal gap and a likely consumer-friction gap. - Agent docs point to a non-existent policy.
llms.txt/agents.mdreference/policies/shipping-policy(404) and/policies/refund-policy. The refund policy actually lives at/pages/returns-cancellations, not/policies/refund-policy(which 404s in this pass's path test). Agent-facing docs and reality are out of sync — fix the docs to point at the real URLs (/pages/returns-cancellations, and create a shipping policy).
6. Current AI Visibility (web-search proxy)
We tested priority queries via web search as a ranking-presence proxy (AI answers draw heavily from pages that rank / are authoritative).
| Query | Provogue cited? | Who gets cited |
|---|---|---|
| "provogue luggage review" | Yes (rank #1) + third-party blogs (mithilacraft, apnabrands, Myntra reviews) | Provogue.com + low-authority review blogs |
| "best hard luggage brands india 2026" | No | GQ, The Planet D, Pinterest, RIMOWA, Samsonite |
| "best luggage brand in india (mokobara nasher safari vip)" | No | Cashaly, Mokobara, NORI, CouponzGuru, ReviewSky |
| Category "cabin luggage" / "printed luggage" / "personalised luggage india" | No organic rank observed | Marketplace + competitor pages |
Conclusion: Provogue is present only on branded queries. On the high-intent consideration queries ("best", "vs", "comparison"), it is invisible, while competitors Mokobara, Safari, VIP, Nasher Miles, Uppercase, and NORI are repeatedly cited in third-party roundups. This is the core AEO problem: AI answers are being written about the category without Provogue in them.
Third-party review blogs that DO mention Provogue (mithilacraft, apnabrands) are thin and low-authority — they neither rank broadly nor lift Provogue into AI answers.
7. Competitive AEO Landscape
Competitors already winning AI "best/comparison" answers:
- New-age D2C: Mokobara (strong own-site + blog), NORI (definitive comparison guides with tables), Uppercase, Nasher Miles.
- Legacy/value: Safari, VIP (American Tourister/Skybags) — cited for trust + value.
- Roundup publishers (Cashaly, ReviewSky, CouponzGuru) aggregate and get cited by Perplexity/ChatGPT.
Provogue's differentiators that AI answers should surface but currently don't: 1997 heritage + "Trusted by Millions", personalised/co-created printed luggage (a genuine moat), 3–5 year warranties, and the Fardeen Khan "OGs Ka Comeback" relaunch. None of these appear in category answers today.
8. Optimization Roadmap (Three Pillars: Structure → Authority → Presence)
Pillar 1 — Structure (make it extractable) [HIGHEST ROI]
| # | Action | Effort | Impact | Why |
|---|---|---|---|---|
| S1 | Add descriptive H1 to collection + about pages (collection title as H1) | Low | High | AI topic signal; currently 0 H1 on these. |
| S2 | Add FAQPage JSON-LD to the 59-question FAQ page | Low | High | Directly extractable by ChatGPT/Perplexity/Claude. |
| S3 | Expand PDP descriptions to 300–800 words (specs, use-case, warranty, comparison bullets) for hero collections first | Med | High | Thin copy is the #1 on-page gap across 233 SKUs. |
| S4 | Add BreadcrumbList schema to collection + PDP templates | Low | Med | Category hierarchy for AI. |
| S5 | Add /pricing.md (structured tiers: luggage/prints/solids/backpacks with INR ranges) | Low | Med | Parseable pricing for agents + humans. |
| S6 | Add gtin / mpn to Product schema where available | Low | Med | Product-ID extraction for comparison queries. |
| S7 | Add hreflang="en-IN" (single-market India) | Low | Med | Geographic targeting signal; currently absent. |
Pillar 2 — Authority (make it citable)
| # | Action | Effort | Impact | Why |
|---|---|---|---|---|
| A1 | Grow review volume (Judge.me / Yotpo) so AggregateRating shows real counts | Med | High | ratingCount:1 undermines trust; also a known conversion gap. |
| A2 | Add author + credentials + "last updated" to blog posts and comparison guides | Low | Med | E-E-A-T; AI weights recency + authorship. |
| A3 | Publish original data (e.g., "India luggage buyer survey", warranty-claim stats) — +37–40% citation boost per Princeton GEO research | Med | High | Unique citable statistics. |
| A4 | Add statistics with dated sources to collection/about pages (e.g., "2M+ customers since 1997", "3–5 yr warranty") | Low | Med | Stats boost non-Google citation. |
Pillar 3 — Presence (be where AI looks)
| # | Action | Effort | Impact | Why |
|---|---|---|---|---|
| P1 | Create comparison/alternative content: "Provogue vs Safari/VIP/Mokobara", "Hard vs soft luggage", "Best cabin luggage India 2026" with comparison tables | Med | High | Comparison articles = ~33% of AI citations; Provogue absent today. |
| P2 | Build a "Where to Buy" page linking D2C + Flipkart + Amazon.in + (future) quick-commerce | Low | Med | Captures high-intent "where to buy" queries; also referenced by agents. |
| P3 | Earn third-party roundup mentions (pitch Cashaly/ReviewSky/CouponzGuru-style publishers; get into NORI/Mokobara comparison posts) | Med | High | 3rd-party cites >> own-domain cites (6.5x per industry data). |
| P4 | Fix agent docs' policy links + publish missing shipping policy page | Low | Med | Trust + agent accuracy. |
| P5 | Keep blog active quarterly with fresh, query-targeted guides (currently Diwali-seasonal/sparse) | Med | Med | Freshness is weighted heavily by AI. |
| P6 | Strengthen Wikipedia / entity presence (heritage "Trusted by Millions" 1997 story) for entity recognition | Med | Med | Wikipedia = 7.8% of ChatGPT citations. |
9. Priority Sequence (do in this order)
- S1 + S2 + S4 + P4 (this week, low effort): H1s, FAQPage schema, BreadcrumbList, fix agent-doc policy links + publish shipping policy. These are cheap and directly improve extractability + agent accuracy.
- S3 + A1 (2–4 weeks): expand PDP copy for hero collections; deploy reviews app. Biggest on-page lift.
- P1 + P3 (1–2 months): publish comparison/alternative content + earn third-party roundup mentions — the lever that actually gets Provogue into AI answers.
- A2/A3, S5/S6/S7, P2/P5/P6 (ongoing): authority content, machine-readable files, presence.
10. Monitoring Plan
| Metric | How | Cadence |
|---|---|---|
| Brand citation rate | Manual: run "provogue luggage", "provogue vs [competitor]" through ChatGPT/Perplexity/Google AI Overviews | Monthly |
| Category citation rate | Same for "best luggage brand india", "cabin luggage india", "personalised luggage" | Monthly |
| Share of AI voice vs competitors | Peec AI / Otterly / ZipTie (cross-platform) | Monthly |
| Referral traffic from AI | GA4 source/medium; watch for chatgpt.com / perplexity.ai / bing referrals | Monthly |
| Schema health | Re-run raw-HTML schema check after template changes | Per release |
Note: Google provides no AI-specific Search Console report — standard Performance/Core Web Vitals still apply for Google; cross-platform AI citation must be tracked with the third-party tools above or manual checks.
11. What NOT to do (per Google's AI optimization guide)
- Do not write separate "AI-only" content or chunk pages into fragments — risk of scaled-content-abuse spam policy. Keep one people-first version (the structural patterns above are normal good organization, not AI-bait).
- Do not block GPTBot / PerplexityBot / ClaudeBot / Google-Extended (currently allowed — keep it that way).
- Do not gate the most authoritative content (comparison guides, warranty, FAQ) behind login/JS walls.
- Do not stuff keywords — keyword stuffing reduces AI visibility (~−10% per Princeton GEO research).
12. Citations / Sources checked
- https://www.provogue.com/robots.txt — AI bot access (all allowed)
- https://www.provogue.com/llms.txt — agent doc (200)
- https://www.provogue.com/agents.md — agent doc (200)
- https://www.provogue.com/.well-known/ucp — UCP merchant profile (200)
- https://www.provogue.com/sitemap_agentic_discovery.xml — agentic sitemap (200)
- https://www.provogue.com/ (raw HTML) — Organization/WebSite schema, H1=logo
- https://www.provogue.com/products/lattitude-core-large-aegean-blue-meadow-yellow (raw HTML) — Product/Offer/AggregateRating(1 review)/Brand
- https://www.provogue.com/collections/luggage (raw HTML) — ItemList, 0 H1, no BreadcrumbList
- https://www.provogue.com/pages/faq-page (raw HTML) — 59 questions, Organization only (no FAQPage)
- https://www.provogue.com/pages/returns-cancellations, /pages/privacy-policy, /pages/terms-conditions-1 — 200 (footer links OK)
- https://www.provogue.com/policies/shipping-policy, /pages/shipping-policy, /pages/shipping, /pages/delivery — 404 (no shipping policy)
- https://www.provogue.com/pricing.md, /pricing.txt — 404 (missing)
- Web-search ranking proxy (2026-08-11): "best hard luggage brands india 2026" → GQ/PlanetD/RIMOWA/Samsonite (no Provogue); "best luggage brand in india mokobara nasher safari vip" → Cashaly/Mokobara/NORI/CouponzGuru/ReviewSky (no Provogue); "provogue luggage review" → Provogue.com #1 + mithilacraft/apnabrands/Myntra reviews
- https://www.provogue.com/ (business context) — from
business-analysis.md: 233 SKUs, 95% mobile, 197/233 on sale, heritage 1997, personalisation moat, 3–5 yr warranty, competitors Safari/VIP/Samsonite/Mokobara/Nasher/Uppercase/NORI - Princeton GEO research (KDD 2024, Perplexity.ai) — citation-method visibility boosts cited in strategy
- Google AI features optimization guide (developers.google.com/search/docs/fundamentals/ai-optimization-guide) — stance on no special markup, no AI-only content
End of report. Companion files in workspace: business-analysis.md, pricing-data.csv, social-media.csv.