Problem
Indian sellers on Amazon.in, Flipkart, and Meesho collectively lose an estimated ₹8,000+ crore annually to fraudulent returns — buyers claiming wrong items, damaged goods, or non-delivery when the seller fulfilled correctly. Filing a dispute manually requires navigating each marketplace's seller portal, attaching evidence, writing policy-aligned appeals, and following up over days — a process most small sellers simply abandon. A seller doing ₹30L monthly GMV can silently bleed ₹2–4L per month in uncontested return fraud.
Solution
SellerShield connects to a seller's accounts via Amazon SP-API, Flipkart Seller API, and Meesho Supplier Portal (with a browser-extension fallback for unsupported endpoints). For v1, the core flow is: (1) nightly scan of returns and payment disputes flagged by anomaly rules — mismatch in item weight, repeat-return buyer IDs, serial number inconsistencies; (2) one-click generation of a dispute letter templated to each marketplace's exact policy language using an LLM; (3) automated submission and status polling with escalation reminders sent to the seller via WhatsApp. The dashboard shows pending disputes, win rate, and cumulative recovered revenue.
Why Now
In March 2026, a16z led a $35M round into Glimpse — a US startup doing the same job for CPG brands on Amazon US — validating that dispute automation is a standalone, high-value SaaS category. India's e-commerce GMV crossed ₹5 lakh crore in FY26 with over 5 million active marketplace sellers; Meesho alone onboarded 700,000 new suppliers in FY26. Amazon SP-API became stable and free for Indian developer accounts in 2024, and Flipkart opened its Seller API to third-party apps in late 2025 — the infrastructure to build this product now exists without scraping.
Target User
First 1,000 customers: sellers with ₹10–50L monthly GMV on Amazon.in or Flipkart, based in Delhi-NCR, Surat, Tirupur, or Mumbai, selling in high-return-rate categories (electronics accessories, apparel, beauty). They typically manage orders via a single laptop and a WhatsApp group with their logistics partner. They know return fraud is happening but have no bandwidth to fight it — the purchase trigger is seeing a monthly P&L where "returns adjustment" is larger than their ad spend.
Business Model
Two tiers billed monthly via Razorpay: Starter at ₹999/month (up to 50 disputes/month, Amazon only) and Growth at ₹2,999/month (unlimited disputes, all three marketplaces) plus 10% of the rupee value of disputes won above ₹10,000 in a month. Gross margin is 82%+ — cost of goods is API call fees (~₹400/month at scale) and LLM tokens (~₹600/month per active account). A 500-seller cohort at blended ₹2,000 ARPU yields ₹10L MRR with a payback period under 2 months given near-zero CAC in seller Facebook groups.
Competitive Landscape
- Direct (India): No pure-play Indian dispute automation tool exists; EcommerceHacks and a few CA-run services offer manual dispute filing as a service, not software.
- Direct (global reference): Glimpse (US, CPG/Amazon US focus, just raised $35M from a16z); Chargebacks911 (card-level disputes, not marketplace return fraud).
- Why we win: India-specific marketplace policy knowledge baked into templates; WhatsApp-native UX matches how Indian sellers operate; pricing 10× cheaper than any US equivalent; Meesho integration that no global player will prioritise.
6-Month Plan
- Month 1–2 (₹3L): Amazon SP-API integration + anomaly detection rules v1 + dispute letter LLM engine; recruit 10 beta sellers from Delhi-NCR seller Facebook groups; validate win rate > 40%.
- Month 2–3 (₹2L): Flipkart Seller API integration; WhatsApp notification layer via Twilio/WATI; browser extension fallback for Meesho; iterate templates based on beta feedback.
- Month 4 (₹2L): Paid launch at ₹999 tier; performance marketing in 5 large seller WhatsApp/Telegram communities; target 50 paying accounts.
- Month 5–6 (₹5L): Growth plan launch with 10% recovery share; Meesho native integration if API opens; ONDC seller connector; hire one sales/support person from the seller community itself.
Risks
- Marketplace API revocation: Amazon or Flipkart could restrict third-party dispute filing via API; mitigated by browser-extension fallback and by framing the product as "assisted filing" rather than automated submission.
- Low win rates eroding trust: If dispute win rates fall below 30%, sellers churn fast; mitigated by building a community-sourced template database that improves with every won case and starting only with high-confidence fraud signals.
- Seller price sensitivity in Meesho tier: Meesho's informal seller base (many first-time e-commerce entrants) may resist ₹999/month; mitigated by a freemium tier (5 disputes/month free) to demonstrate ROI before asking for payment.
Score Breakdown
Market (15/20): Five million-plus active Indian marketplace sellers represent a reachable TAM of ₹3,000 Cr+ at even 10% penetration and ₹2,000 ARPU — well above the ₹1,000 Cr threshold, but scored 15 not 20 because the immediate serviceable market is the subset of mid-GMV sellers who are aware enough to pay for tooling.
Capital (12/15): The entire stack is pure software — SP-API calls, an LLM prompt layer, a React dashboard, and WhatsApp webhooks; MVP builds comfortably within ₹6–8L, leaving ₹4–6L for 6 months of cloud and marketing, well under the ₹20L cap. Scored 12 not 15 because LLM token costs could spike if dispute volume scales faster than revenue.
Team (8/10): Two backend engineers (one with API/webhook experience, one with LLM prompt engineering) plus one founder doing sales; v1 is achievable in 8 weeks. Scored 8 because Flipkart's seller API has sparse documentation and one engineer will spend 2+ weeks on integration edge cases.
Trend (13/15): A direct US analog — Glimpse — just raised $35M from a16z in March 2026 for the identical product category on Amazon US; India's marketplace GMV and seller base are at an inflection point. Scored 13 not 15 because the Indian signal is extrapolated from the US raise rather than a native Indian traction data point.
Moat (10/15): Proprietary dispute template database trained on thousands of won/lost cases becomes a compounding data moat; marketplace API credentials create switching cost once embedded in a seller's workflow. Scored 10 not 15 because templates are copyable if a well-funded competitor enters and marketplace policy changes can reset the template library overnight.
Economics (13/15): 82%+ gross margin, near-zero marginal cost per dispute filed, and a recovery-share upside component that aligns incentives; CAC is near-zero in seller communities. Scored 13 not 15 because the 10% recovery share introduces revenue volatility and some sellers will negotiate it away at scale.
Speed (7/10): Amazon SP-API is well-documented and stable; first paying user reachable in 8 weeks via a focused beta. Scored 7 not 10 because Flipkart and Meesho integrations add complexity and the full multi-marketplace product takes 14–16 weeks.