CreditWarden

1. One-liner

Ranks your monthly GST IMS invoice inbox by risk and tells you exactly what to accept, reject, or hold before the 14th.

2. Trend signal — why now?

Two regulatory changes landed on top of each other in 2026, and they turned a once-optional reconciliation into a monthly, money-losing chore for every GST-registered business in India.

• IMS became mandatory on 1 April 2026. The Invoice Management System now sits between every supplier’s GSTR-1 and your GSTR-2B. You must Accept / Reject / keep-Pending each inward invoice before the 14th, or the system marks it deemed accepted and pulls it into your credit. Tax Garden: “From April 1, 2026, every GST-registered business must use IMS to manage inward supplies before filing GSTR-3B… For small shops, kirana stores, wholesalers, and MSME owners across India, the compliance burden just got real.”
• GSTR-3B ITC hard-locking is targeted for ~July 2026. After it lands, ITC in GSTR-3B auto-populates from GSTR-2B with no manual editing. Taxscan/CAclubindia: “every mismatch between your purchase register and GSTR-2B will become a locked-in error you cannot fix inside GSTR-3B.” The escape hatch — reject the bad invoice before you file — only works if you catch it first.
• The buyer eats the supplier’s mistakes. CAclubindia titled a whole piece “Input Tax Credit Under GST: The One Provision Where a Buyer Pays for a Supplier’s Default” — “If the supplier files late, files wrong, or does not file at all, the buyer’s credit simply does not show up,” and misclassifying a blocked credit “does not just invite reversal — it attracts interest under Section 50.”

The behaviour this produces is exactly the wedge. TaxGuru’s IMS piece: “Most businesses hurt themselves by accepting every invoice without reviewing it (claiming ITC they should not), or leaving everything pending and then scrambling the night before GSTR-3B is due.” “No Action is an Action.” Nobody owns the IMS inbox, so it either gets rubber-stamped (illegal ITC, future notice) or ignored (lost ITC).

Provenance:

• Signal 1 (Demand): IMS mandatory Apr 2026; MSMEs/kiranas now must action every inward invoice monthly, “compliance burden just got real” — https://taxgarden.in/blog/ims-invoice-management-system-mandatory-gst-2026 — 2026-07-04
• Signal 2 (Feasibility/Economic): GSTR-3B ITC hard-locking ~July 2026 turns every 2B-vs-books mismatch into a locked, interest-bearing error — https://www.taxscan.in/top-stories/big-gst-change-from-july-2026-gstr-3b-itc-locking-explained-1448389 — 2026-07-04
• Signal 3 (Demand/Economic): Buyer pays for supplier default; reversal + Section 50 interest; heavy manual reconciliation burden — https://www.caclubindia.com/articles/input-tax-credit-under-gst-the-one-provision-where-a-buyer-pays-for-a-supplier-s-default-55832.asp — 2026-07-04 Category: Regulatory arbitrage

3. The opportunity

The incumbents — ClearTax, Zoho Books, TallyPrime, IRIS, Cygnet — all now have an “IMS reconciliation” screen. But they are full GST/accounting suites: ₹10,000/yr+, sold to the CA firm or the finance team, requiring you to import your purchase register, map ledgers, and live inside their product. Their IMS module is a table of 500 invoices with Accept/Reject checkboxes bolted onto a spreadsheet. It answers “here are all your invoices.” It does not answer the only question the owner actually has: “which of these 40 invoices should I be worried about, and what do I do about each one, before the 14th?”

That gap is the product. CreditWarden is a single-purpose IMS action-desk that sits alongside whatever accounting software the business already runs. It pulls the IMS/GSTR-2B feed, ranks every invoice by risk (new supplier, amount doesn’t match your books, GSTIN that hasn’t been filing, duplicate, blocked-credit category), and hands you a triaged worklist with a recommended action and a one-line reason for each. Then it nags you — on WhatsApp — before the deadline. It is the judgment layer the suites skip, priced for the wallet the suites price out.

10× better than the incumbent screen on the one job that matters: turning a 300-row invoice table into a 10-minute, 8-decision task with the risky ones on top.

4. Target market

• Primary customer: Small GST-registered businesses — traders, wholesalers, distributors, small manufacturers, D2C sellers — with roughly ₹1–20 crore turnover and 30–800 inward invoices a month. Also the small CA / tax-practitioner firm (2–15 staff) that files GSTR-3B for 40–300 such clients and now has to touch every client’s IMS inbox every month.
• Why they buy (in their words): “Reconciliation takes hours.” “Invoices pile up.” “We either accept everything and get a notice, or we scramble the night before 3B.” The CA’s version: IMS just added a whole new monthly step per client with no extra fee they can charge for it.
• Rough TAM reasoning: India has well over 1.4 crore active GST registrations. Even the sliver that (a) has enough inward invoices to feel IMS pain and (b) will pay for a tool is in the low millions of businesses, plus tens of thousands of small CA firms. A niche of even 50,000 paying accounts at ₹6–12K ACV is a ₹30–60 crore business — far past our $5M ceiling.
• Why now for them: Before April 2026 IMS was optional and most ignored it. The hard-lock (~July 2026) removes the “fix it later in 3B” safety net. The pain went from theoretical to monthly-and-metered in one quarter.

5. Product sketch (MVP)

• One-tap IMS pull — connect via GSP/GST-portal API (or upload the IMS/2B export), no full accounting migration required.
• Risk-ranked worklist — every inward invoice sorted by a risk score, not by date. Riskiest at top: new/unverified GSTIN, supplier who hasn’t filed, amount ≠ your purchase entry, duplicate, blocked-credit HSN, first-time vendor.
• Recommended action per invoice — Accept / Reject / Pending with a plain-English one-liner (“Reject: supplier GSTIN cancelled last month” / “Hold: ₹18,400 in 2B vs ₹18,000 in your books”).
• Books-vs-2B match — optional lightweight import of your purchase register (Tally/Zoho/Excel) to flag mismatches and missing invoices (supplier didn’t file → your ITC is at risk).
• Deadline warden — WhatsApp + email countdown to the 14th; “22 invoices still un-actioned, 3 flagged high-risk.”
• Deemed-acceptance guard — a pre-14th “you’re about to auto-accept these 3 risky invoices” alert.
• Filing-ready summary — a clean before-you-file recap: eligible ITC, rejected/held items, and the reasons, exportable for the CA’s file.
• Multi-client console (for CA firms) — one dashboard across all client GSTINs, sorted by whose inbox is most on fire.

6. AI angle — what’s load-bearing

AI does the triage judgment, which is the whole product. Rules alone can flag GSTIN cancellations and exact amount mismatches, but the hard part is ranking ambiguous cases the way an experienced accountant would: “this supplier’s invoice number pattern is off,” “the description doesn’t match what this vendor usually bills you,” “this looks like the same invoice you already have under a different number.” An LLM layer classifies invoices, reconciles fuzzy line-item descriptions against the purchase register, and drafts the one-line “why” and recommended action for each. Remove the AI and you’re back to the incumbent’s dumb 500-row table — which already exists and which nobody wants to work through. The AI is what turns 300 rows into 8 decisions.

7. Localization angle

This is intrinsically India-first — it exists because of Indian GST/IMS mechanics; there is no global version. The localization is the moat:

• Deep GST rule knowledge (Section 16/17 blocked credits, Rule 37 180-day reversal, Section 50 interest, IMS deemed-acceptance timing).
• WhatsApp-first delivery — the owner and the CA’s article clerk both live on WhatsApp, not email.
• ₹499–1,999/mo pricing that works where a $49/mo tool cannot.
• GSP/GST-portal API integration and the 14th-of-month cadence baked in.

8. Business model — path to $1M–$5M ARR

• Pricing:
  • Micro (single GSTIN, <100 invoices/mo): ₹499/mo
  • Business (single GSTIN, higher volume + books-match): ₹1,499/mo
  • CA firm (per-client seats, multi-GSTIN console): ₹5,000–25,000/mo depending on client count
• ACV: ~₹9,000 blended (mix of ₹6K single-business and ₹60K+ CA-firm accounts).
• To $1M ARR (~₹8.3 cr): ~9,000 single-business accounts at ₹6K, or ~1,400 CA firms at ₹60K, or (realistically) a blend of ~3,000 businesses + ~700 CA firms.
• To $5M ARR: land the CA channel hard — each CA firm resells the seat to their clients, so ~3,500 CA firms averaging ~₹1.2L ACV gets there. CA firms are the leverage; direct SMB is the top of funnel and the credibility proof.
• Expansion path: start at IMS triage → add books-match → add GSTR-1 outward reconciliation → add vendor-follow-up automation (auto-WhatsApp the defaulting supplier) → per-GSTIN seat growth inside CA firms.

9. Go-to-market wedge — first 100 customers

• CA/tax-practitioner WhatsApp & Telegram groups. There are hundreds of large regional GST-practitioner groups where IMS/hard-lock is the anxiety of the month. Drop a genuinely useful “IMS risk-triage” free-tier link when the topic comes up; these groups pass tools around fast. Target: 40 CA-firm trials from 6 weeks of participation.
• Free “IMS Risk Scan.” A no-signup tool: upload your IMS export, get a ranked risk report + count of high-risk invoices, PDF branded. It’s the demo and the lead magnet. Push it via the groups above and via the tax-content YouTubers/LinkedIn CAs who explain GST changes to SMBs.
• Co-sell with 3–5 mid-tier GSPs / billing tools that have IMS data but a weak IMS UX — bundle CreditWarden as the “smart IMS” add-on their SMB users already need.
• Direct to distributors/wholesalers in 2–3 high-invoice-volume mandi/trade clusters (auto parts, pharma distribution, FMCG) where 200+ inward invoices/month is normal — cold WhatsApp with a pre-run risk report on their own GSTIN once they connect.
• The first 100 come mostly through the CA channel: 30–50 CA firms, each with a handful of their most IMS-stressed clients.

10. Build complexity — justification

Medium. The moving parts are a GSP/GST-portal API integration (IMS + GSTR-2B pull — real but well-trodden; several GSPs offer sandboxed APIs), a risk-scoring engine (rules + an LLM classification/reconciliation layer, all off-the-shelf model APIs), lightweight purchase-register import, and WhatsApp Business API for alerts. No custom models, no novel infra. The gnarly part is GST-rule correctness and staying current with GSTN advisories — that’s domain work, not engineering research. A pair (one strong engineer + one GST-fluent operator/CA) ships a credible v1 in ~10–14 weeks.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Helps taxpayers comply; reads their own GST data via authorized GSP APIs.
Ethical — no harm / dark patterns	✅	Reduces both illegal over-claiming and accidental ITC loss. Pro-compliance.
Market exists (evidence above)	✅	Mandatory IMS + hard-lock + documented reconciliation pain.
1–5 person team can build this	✅	Pair ships v1 in ~3 months.
Launchable with <$50K / ₹40L	✅	GSP API fees + model/WhatsApp API + two builders.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Monthly, deadline-bound, money-losing on both sides (lost ITC / illegal ITC + interest). Not quite daily hair-on-fire, but recurring and metered.
Demand evidence	15	12/15	Two hard regulatory triggers + documented CA/MSME complaints + a funded GST-software category. Docked because most “pain” quotes are from vendors, not raw customer forums.
Build feasibility	15	11/15	Off-the-shelf stack, but GSP API integration + GST-rule correctness need care. Not a weekend build.
Distribution clarity	15	12/15	CA WhatsApp/Telegram groups are a real, named, fast channel with a natural free-scan lead magnet. Conversion still to be proven.
Revenue mechanics	15	12/15	Priced against ₹10K+/yr incumbents; CA-firm reseller motion gives real ACV leverage. Micro tier margins thin.
Time to first revenue	10	8/10	Free scan → paid in the same monthly cycle; a CA firm can convert in weeks. Pre-14th urgency shortens the sale.
Defensibility	10	6/10	Soft moat: GST-rule depth, CA-firm workflow lock-in, accumulating risk-pattern data. Incumbents can copy the feature; the focus + channel is the edge.
Total	100	77/100

13. Qualitative modifiers

Founder-fit tags

domain-expertise-required · technical-heavy — needs a GST-fluent founder (or a CA co-founder/advisor who lives the pain) paired with an engineer who can handle GSP APIs and the model layer.

Key assumptions to validate (3–5)

1. Assumption: Small businesses and CA firms will pay ₹499–₹5,000/mo for a judgment/triage layer on top of IMS data they can technically see for free on the portal. How to test: Put the free risk-scan in front of 30 CA firms; count how many ask “can it just auto-tell me what to do each month” and would pay for it.
2. Assumption: GSP/GST-portal API access reliably exposes the IMS + 2B data we need at a per-account cost that leaves margin at ₹499/mo. How to test: Sign a GSP sandbox agreement in week 1–2 and price the per-GSTIN API cost before building.
3. Assumption: The CA firm is the real buyer and will resell seats to clients. How to test: Get 5 CA firms to run 3 client GSTINs each through a manual-service pilot and see if they’d charge clients for it.
4. Assumption: The AI triage is meaningfully better than incumbent rule-tables at ranking risk. How to test: Blind-test our ranked worklist vs a ClearTax/Tally IMS screen with 5 accountants on the same month’s data.

Risk flags

1. Platform dependency: Total reliance on GSTN/GSP APIs and on GSTN’s own rules. If GSTN improves the native IMS UX or adds risk-flagging, the wedge narrows. Mitigate by moving up the value chain (vendor follow-up, CA multi-client ops) fast.
2. Incumbent reaction: ClearTax/Tally/Zoho already have IMS screens and the customers. If one ships good risk-triage, we’re fighting on their turf. Our defense is focus, price, and the CA-channel motion — not features alone.
3. Regulatory timing: The hard-lock date has slipped before. If it slips again past mid-2026, urgency (and the “you can’t fix it later” pitch) softens for a few months. The IMS-mandatory reality remains regardless.

14. Structured verdict

Score:                  77/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       GST-fluent founder / CA co-founder + one strong full-stack engineer
Time to revenue:        6–10 weeks (free scan → paid within a monthly IMS cycle)
Capital to launch:      ₹6–12 lakh ($7–14K) — GSP API fees, model + WhatsApp API, two builders
Top 3 assumptions to validate first:
  1. Willingness to pay for a triage/judgment layer over free portal data — 30 CA-firm free-scan test
  2. GSP API exposes IMS+2B at a cost that leaves margin at ₹499/mo — sandbox agreement week 1–2
  3. CA firm is the buyer and will resell seats — 5-firm manual-service pilot
Kill criteria:
  - Abandon if <15% of 40 CA firms running the free scan ask to keep it / pay after one cycle
  - Abandon if GSP per-GSTIN API cost makes the ₹499 micro tier unprofitable and no CA-firm bundle closes
  - Abandon if a top-3 incumbent ships equivalent risk-triage before our v1 and bundles it free

15. Next step — 1-week validation sprint

• Day 1–2: Sign a GSP sandbox NDA; confirm the IMS + GSTR-2B API surface and per-GSTIN cost. In parallel, build a throwaway “IMS Risk Scan” that takes a real IMS/2B Excel export and outputs a ranked risk report (rules + one LLM pass) — no login, no product.
• Day 3–4: Drop the free scan into 3–4 large CA/GST-practitioner WhatsApp/Telegram groups and DM 20 small CA firms directly. Ask each to run one real client month through it.
• Day 5: Decide go/no-go. Green light if ≥15 firms run real data through it and ≥6 say, unprompted, “can this just run every month / tell me what to do — I’d pay for it.” Falsifiable: if the ranked worklist doesn’t beat their current portal/Tally workflow in their own words, it’s a no-go.